Citation
Site-based mating system in a tropical harvestman

Material Information

Title:
Site-based mating system in a tropical harvestman
Creator:
Mora, Giselle, 1958-
Publication Date:
Language:
English
Physical Description:
viii, 190 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Animal nesting ( jstor )
Breeding seasons ( jstor )
Eggs ( jstor )
Female animals ( jstor )
Fungi ( jstor )
Hatching ( jstor )
Juveniles ( jstor )
Mating behavior ( jstor )
Mating systems ( jstor )
Nesting sites ( jstor )
Dissertations, Academic -- Zoology -- UF
Zoology thesis Ph. D
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 181-188).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Giselle Mora.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
027082202 ( ALEPH )
25622382 ( OCLC )

Downloads

This item has the following downloads:


Full Text












SITE-BASED MATING SYSTEM IN A TROPICAL HARVESTMAN
















BY


GISELLE MORA

















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA 1991














ACKNOWLEDGEMENTS



I have benefited during the writing of this thesis and my graduate studies from the encouragement, comments and criticisms of many people. It has been with their help that I was able to successfully complete this project. I would like to thank the members of my committee, H. Jane Brockmann, Jonathan Reiskind, John Sivinski, John Anderson and William Eberhard for their valuable comments throughout the various stages of my research and for their criticisms during the final revision of this work. Special thanks go to my adviser H. Jane Brockmann. She has contributed immensely to organizing the material, to improving the quality of this thesis and in many other aspects of my education. Linda Fink has also contributed with ideas, comments and has been a supportive friend throughout. Julio Arias helped with statistical analyses.

The Department of Zoology, University of Florida, has been a stimulating and supportive place for conducting my graduate studies. I would like to thank Dr. Frank Nordlie, Mrs. Lynda Everitt and Mrs. Carol Binello for logistic support.

I have also received much support and encouragement while at the Smithsonian Tropical Research Institute in



ii










PanamA. Drs. Egbert Leigh and Donald Windsor acted as Smithsonian advisors of this project. Egbert Leigh has supported this research since 1983 and gave me the confidence for continuing during the darkest moments of my field work. I thank him for not letting me forget the value of natural history. H4ctor Barrios assisted me in the field during the 1987 season. Argelis RomAn and Georgina de Alba provided logistic support and help with many matters during my stay in PanamA. Rayneldo Urrutia coordinated many administrative aspects of my residence at BCI. I overlapped with many researchers during my residency at BCI; many of them contributed with observations, criticisms and friendship. In special, I appreciate the encouragement I always received from Phil DeVries.

The support of my family and friends have brightened my years as a student. I thank my entire family for their support and love. In special, I thank my parents for always letting me try to be all what I thought I could be. I also thank Juan Carlos Vargas, Julio Arias, Clara Sotelo, and Catherine Langtimm for their friendship and support especially during the writing of this thesis.

My husband Henry put up with me, fed me, and kept his and my spirits high when the writing of this thesis got rough. I thank him for never losing his sense of humor and for helping me finish.

I have been supported by a Noyes Predoctoral Fellowship from the Smithsonian Tropical Research Institute, by teaching iii










and research assistantships from the Department of Zoology, University of Florida and by a scholarship from the Consejo Nacional para Investigaciones Cientificas y Tecnol6gicas, Costa Rica.

















































iv















TABLE OF CONTENTS



ACKNOWLEDGEMENTS ........................................... ii

ABSTRACT .................................................. vii

INTRODUCTION ................................................1

BACKGROUND AND METHODS ......................................5

Natural History and Reproductive Biology of Opiliones.....5 Reproductive Biology of Zvaopachvlus albomarainis.........12 Study Site................................................14
Study Population.........................................16
General Methodology......................................16
Definitions..............................................18

NESTS AND MALE SUCCESS .....................................22

Nests and Male Success: General Descriptions.............23
Patterns of Nest Use.....................................28
Patterns of Nest Success.................................32
Male Life History and Behavior: General Descriptions.....39 Components of Male Success ...............................43
Why do Males Abandon Nests?..............................50
Nests or Males? A Male Switching Experiment ............. 52
Discussion: Patterns of Nest and Male Success............55

FEMALE ASSOCIATIONS WITH NESTS AND FEMALE SUCCESS .......... 80

Types of Visits and Female Condition ..................... 81
Patterns of Female Associations ..........................87
Female Associations and Nest/Male Measurements...........97
Components of Female Success .............................99
Discussion: Females Association with Nests............... 103

DO FEMALES ASSOCIATE WITH MALES OR NESTS? ................. 120

Methods.. ............................................... 121
Results: Female Residency Before and After the
Switchings ...........................................123
Discussion: Females Associations With Nests............123






v










THE NEST SITES ............................................ 127

Rainfall................................................129
Fungus..................................................138
Predators.................... .........143
What Accounts for the Differences in Nest Success? ...... 146 SITE-BASED MATING SYSTEM OF Zvaonachylus albomarainis ..... 162

The Mating System of Z. albomarainis .................... 165
The Evolution of the Site-based Mating System of Z.
albomarainis. ........................................ 176

LITERATURE CITED ..........................................181

BIOGRAPHICAL SKETCH ....................................... 189








































vi














Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


SITE-BASED MATING SYSTEM IN A TROPICAL HARVESTMAN

By

Giselle Mora

December, 1991




Chairperson: H. Jane Brockmann Cochair: Jonathan Reiskind
Major Department: Zoology



ZvaoDachylus albomarainis Chamberlin (Arachnida,

Opiliones: Gonyleptidae), a tropical harvestman, is the only arachnid exhibiting paternal care. Males construct mud nests on trees that females visit to court the resident male and oviposit. I conducted a two-year field study on Barro Colorado Island, Panama to identify and evaluate components of reproductive success and to characterize the mating system of a natural population of this species. The mating system of these harvestmen is polygynandrous, with females laying eggs in several male nests and males collecting eggs from several females. Females mature a few eggs at a time throughout a breeding season lasting from June to December. Females exhibited a random pattern of nest visitation but the



vii









distribution of ovipositions was not random. This mating system resembles those of many substrate-nesting, eggguarding fish. It is site-based since both male and female reproductive success depends upon their association with successful nesting sites. Breeding success of males was measured by (a) the number of eggs acquired during his residency (b) the number of juveniles hatched, and (c) the number of resident females associated with the male and his nest. The ability of males to acquire and hold a nest at a good nesting site accounted for much of the variability in male breeding success. Male reproductive success was not correlated with body size. Switching experiments revealed that resident females continued to associate with the site regardless of the identity of the male at that site. Ecological variables, including rain, occurrence of fungal attack and incidence of predation may account for the success of some nesting sites over others. I hypothesize that this mating system evolved from no parental care. A limited availability of safe oviposition sites may account for the evolution of male territoriality, male egg-guarding and female attraction and fidelity to nesting sites.













viii














CHAPTER 1
INTRODUCTION



The harvestman Zygopachylus albomarainis (Opiliones:

Gonyleptidae) is the only arachnid exhibiting paternal care (Mora 1990). Their reproductive behavior is remarkable for an arachnid. Males guard eggs in nests made out of mud and tree bark. Besides building a nest, males take over or use abandoned nests. Females visit nests, court males, oviposit, and leave the eggs for the male to guard. These behaviors are unique among opilionids.

I present here the results of a field study that focuses on the variation in reproductive success of males and females and the factors that contribute to such variation. The objectives of the study were to identify and evaluate components of reproductive success and to characterize the mating system of a population from Barro Colorado Island, Panama. My ultimate goal was to gain understanding of selective pressures that have shaped this mating system. The importance of this study lies in documenting the reproductive behavior of the only arachnid with paternal care. It also allows a test of current ideas on the influence of increased paternal investment on male and female mate choice and relative variation in reproductive success.




1






2



The first model explaining the factors contributing to the variation in reproductive success was proposed by Darwin (1871). He recognized two types of selective forces acting on the context of reproduction. Intrasexual selection is the competition within one sex for individuals of the opposite sex and intersexual selection is the preferential choice for mating partners by one sex. Darwin considered intrasexual selection to be more common and more pronounced in males and intersexual selection to be an attribute of females. Although this distinction has proven useful, most biologists now recognize the difficulty of separating the two (Maynard Smith 1987).

Trivers (1972) elaborated Darwin's idea by arguing that the sex with the greatest parental investment will become a limited resource for the opposite sex. He recognized as parental investment those activities of an animal that increase the probabilities of present offspring survival and reproduction at the expense of the parent's ability to have additional offspring. Consequently, the sex investing the least will compete among themselves in order to mate with individuals of the sex investing the most. Since females generally produce fewer and energetically more expensive gametes than males, their parental investment is often greater. This produces the general pattern of male-male competition and female mate choice. When a male invests, he may become a limited resource for females and the pattern may be reversed. Also, phylogenetic and ecological factors will






3



produce variation in the degree of female competition for males and male choosiness. Such factors have been investigated for some species (Petrie 1983), but they have not been evaluated in any harvestman.

Evaluating the factors that contribute to the variation in reproductive success of males and females is necessary to understand any mating system. For Z. albomarginis males, among the factors that may contribute to individual variation in reproductive success are age, size, and the ability to acquire and retain a nest. Since egg hatching success depends on the presence of a male (Mora 1990), differences in paternal behavior may account for variance in reproductive success. Paternal behavior may also be a criterion by which females assess male quality (Petrie 1983). Female reproductive success is associated with the success of the nest in which she oviposits. Little was known about the behavioral ecology of female reproduction prior to this study. Physiological condition and attachment to nesting sites will be explored as contributors to the variation in female reproductive success. These attributes and behaviors of males and females have been recognized as important contributors to reproductive success in many species (Clutton-Brock 1988) but have never been evaluated for this harvestman.

Since males without nests do not acquire mates, I evaluate in Chapter 3 the patterns of nest use and nest success, and present experimental evidence showing the





4



contribution of nests to the male's reproductive success. For females, an important reproductive choice is whether to associate with a male and his nest or to search for additional mates. I discuss the nature and the patterns of these associations in Chapter 4. Whether females associate with males or with nests was tested with a male switching experiment; I present these results in Chapter 5. In addition, I discuss the ecological setting of the nesting sites in Chapter 6, and characterize the mating system and the components of male and female reproductive success in Chapter 7. I argue that this is a nest site-based mating system; that is, a system in which both male and female behavior and reproductive success depend upon their association with good nesting sites. Finally, I discuss a possible scenario for the evolution of this mating system in Chapter 7.














CHAPTER 2
BACKGROUND AND METHODS



Natural History and Reproductive Bioloay of Opiliones


Reproduction in Zygopachylus albomarainis Chamberlin includes a number of behaviors unusual or unique among Arthropoda. However, this apparent exceptionality may be simply due to our lack of knowledge of natural history of many arthropod groups. Despite the fact that the arthropods constitute approximately four-fifths of the world's animals (Savory 1977), most are poorly known. Our ignorance is particularly evident in the Arachnida, a large and diverse class composed of 12 living and five extinct orders. Of these, the Araneae (spiders) and Acari (mites and ticks) are the best studied. Field and laboratory studies on the ecology and behavior of the rest of the arachnid groups are practically nonexistent and limited mostly to temperate species. Most studies on opiliones have been on species of the suborder Palpatores perhaps because these are common in North America and Europe where a longer tradition of ecological studies exists (for a review see Savory 1977; Hillyard and Sankey 1989).

After Acari and Araneae, the Opiliones are the third

largest order within the Arachnida. They are distinguished


5





6



from other arthropods by having a body divided into two parts (an anterior cephalothorax, prosoma, and a posterior abdomen, opisthosoma); a pair of chelicerae, a pair of pedipalps and four pairs of legs, no antennae or mandibles, and simple eyes. The order is also characterized by the presence of a pair of odoriferous or repugnatorial glands that are usually situated at the bases of the first or second pairs of legs. The cuticle of the body and appendages in opiliones present a range of hairs, setae, spines, denticles, and tubercles that form the "armature". These projections can sometimes become very elaborate (Rambla 1975; Hillyard and Sankey 1989). Opiliones, unlike other arachnids, respire by means of tracheae.

The group is estimated to contain between 3500 and 5000 species (Rambla 1975; Hillyard and Sankey 1989). Species of harvestmen generally do not have an extensive distribution and have little capacity for long range dispersal (Savory 1977). Three suborders are presently recognized (Hillyard and Sankey 1989): the Cyphophthalmi is a very small and primitive group with approximately 50 known species; members of this group resemble mites. The Laniatores or short-legged harvestman are characterized by having the coxae of the first three pairs of legs meeting at the midline, and the pedipalps having a characteristic large terminal claw. It is the dominant group of harvestmen in the tropics and subtropics. Cosmetidae and Gonyleptidae are the largest families and are most diverse in South and Central America where they





7



constitute one-fourth of the suborder; Z. albomarainis is a laniatorid gonyleptid. The suborder Palpatores are the very familiar daddy-long-legs that are common in the north temperate zone; they are recognized because the coxae of the first three pairs of legs do not meet at the mid-line. Phylogenetically, Opiliones are most closely related to Acari; notostigmatic mites are believed to be very closely related to cyphophthalmid opilionids (Savory 1977). Natural History


Like many other arachnids, opilionids are long lived. Juberthie (1967) recorded a cycle of nine years for a cyphophthalmid. However, most of the temperate species have an annual life cycle (Edgar and Yuan 1968) overwintering as near-adults or as eggs (Hillyard and Sankey 1989). Most tropical species probably have life cycles of more than a year with subsequent overlapping of generations (personal observation). It may take up to eight molts to attain adulthood.

The ecology of the early stages of opilionids is a

mystery for most species. In some Leiobunum species, young individuals ascend to the tree and shrub canopy as they mature (Edgar 1971). Tropical species may spend juvenile phases in the forest litter or in the canopy as revealed by evaluation of museum collections (personal observation). Adult individuals commonly form large aggregations, familiar





8



to all, but little studied and understood (Coddington et al., 1990). There have also been observations of territoriality with male-male antagonistic behavior reported for some species of Leiobunum (Edgar 1971).

Because of their secretive lives and inconspicuous appearance, they have few known important predators. Nevertheless, predation by frogs, lizards, birds, mammals, centipedes, spiders and several insects has been reported (Cloudsley- Thompson 1968).

Most species are omnivores, and opilionids have been recorded feeding on other harvestmen, snails, worms, millipedes, woodlice, earwigs, flies, bodies of dead ants and beetles. In addition, they nibble on gills of fungi, dead moles and mice, and bruised and fallen fruit (CloudsleyThompson 1968; Savory 1977). The diet of Opiliones differs from that of spiders. Although the latter are limited to predigested prey juices, the former chew on soft food. Cannibalism, especially among juveniles, is common (Edgar 1971).

Changes in temperature and humidity have been suggested as the major factors of mortality for the group. They are more susceptible to dehydration than most other arachnids. In the laboratory, harvestmen have great difficulty molting when the temperature and humidity are not adequate (Hillyard and Sankey 1989).

Very little is known about the behavior of natural

populations of opilionids. Most opilionids are nocturnal.





9



The standard behaviors reported for harvestmen are negative phototaxis (movement away from light), thigmotaxis (tendency to press the body against the substrate), hydrotaxis (sensitivity to presence of water), and rheotaxis (response to movements). They can autotomize legs when captured, exert secretions from repugnatorial glands, and show catalepsy (faking death) when disturbed (Hillyard and Sankey 1989). Reproductive Bioloav


Opilionid reproductive biology, like that of the Acari, is unusual among Arachnida in their direct use of an intromittent organ during copulation (Berland 1949; Thomas and Zeh 1984). Copulation, according to most observers, "appears to be a casual affair occurring freely and frequently" (Hillyard and Sankey 1989). Studies of tropical laniatores may prove this statement wrong. Reproduction in Cyphophthalmi involves the production of a spermatophore; therefore, copulation in this group does not occur as in other opilionids (Hillyard and Sankey 1989).

Generally there is no marked sexual dimorphism in opilionids; however, in some species there are notable differences in palpi and abdominal sizes (Thomas and Zeh 1984; Hillyard and Sankey 1989). The reproductive system of the females is composed of a horseshoe-shaped ovary, with its anterior ends joined to a median oviduct that leads to the ovipositor which is a broad and mobile tube made up of





10



membranous rings. The reproductive system of males is a Utube shaped testis, with a median vas deferens leading to the penis which is more rigid than the ovipositor. Penis morphology is a species-specific character (Berland 1949; Savory 1977). Both ovipositor and penis lie behind the genital operculum, a fold of tissue that covers the reproductive organs when at rest. The genital opening, although it is derived from abdominal segments, is located between the leg coxae.

During copulation, partners face each other while the male grasps the female with his pedipalps and extends the second pair of legs outwards. The penis is then extruded and the female may use her pedipalps and chelicerae to guide it to the genital opening (Juberthie 1965). The duration of copulation can vary from a few seconds to several minutes depending on the species and the behavior of the female. Repeated copulation with a female seems to be the rule for the palpatorid species for which information exists (Edgar 1971; Hillyard and Sankey 1989). The only existing detailed description of courtship and copulation for a laniatorid comes from Z albomarainis (Mora 1987). That study reports a much more complicated courtship and copulatory behavior than previously thought. Generally, the female approaches the nest, enters and taps the substrate. She then alternates, in variable order, between tapping the partner, moving around the nest, facing the partner and remaining quiscent until the male starts tapping the female. He then moves to her side,





11



her back, and underneath her as he taps the substrate. This continues until the partners face each other and the female extends her chelicerae and palps, grabbing the male by the cephalothorax and pulling him toward her. In this position, the male everts his penis and intromission takes place. Oviposition always follows within 20 min (Mora 1987).

Parthenogenesis may be common in opilionids. For some species, males are extremely rare in collections and natural populations. Parthenogenesis has been documented in Phalanaium opilio and Megabunus diadema (Phillipson 1959; Edgar 1971). To further test parthenogenetic capabilities of M. diadema, Phillipson (1959) collected 10 subadult females and kept them in aquaria without males; seven of them produced 13 batches of eggs, all viable.

Fertilization occurs during the process of egg laying when the sperm are released from the spermathecae, enabling one sperm cell to enter each egg before the hardening of the chorion and vitelline membrane (Juberthie and Muhoz-Cuevas 1971; Juberthie and Manier 1976). In nature, opilionid eggs are difficult to find and very susceptible to the attack of mold (Edgar 1971; Savory 1977). They are usually spherical and of pale coloration. Generally, they are laid in clusters in places such as damp soil, moist vegetable debris, crevices of bark, under stones or dead wood (Rambla 1975; Hillyard and Sankey 1989). In the 10 British species examined by Phillipson, the number of eggs laid was always only a fraction (21-57%) of the mean maximum number of eggs produced





12



by the ovaries. This seems to be a common feature of other species (personal observation).


Reproductive Biology of Zygopachylus albomarginis


The reproductive biology of Z. albomarginis is

extraordinary among the arachnids. This very small and cryptic species (body approximately 3.5 mm long) is known to occur only in Panama. There is no sexual dimorphism in this species. Females are a little larger than males, but the differences are not completely reliable for identification purposes. The species was redescribed from specimens collected on island by Goodnight and Goodnight (1942). The first report on the nesting behavior of this species was provided by Rodriguez and Guerrero (1976). Perhaps the most remarkable aspects of this system are the construction of nests by males and the existence of paternal care. No other arachnid is known to have male parental care, although female care is common. Nest construction in the arachnids is found in spiders (e.g. Salticidae, Pisauridae) and in the brooding nests of pseudoscorpions, though it is the female who guards eggs in those cases (Savory 1977). The structure constructed by males of Z. albomarginis, utilizing mud, tree bark, and salivary secretions is unique among arachnids and its evolutionary origin is a mystery.

Paternal care in this species takes the form of egg guarding inside circular, open mud nests which are





13



constructed on standing or falling trees. Males guard the eggs until they hatch, which takes 18-22 days, preventing the growth of fungus and predation by conspecifics and ants. They also repair the nests when damaged by rain, wind, or the passing of larger animals (Mora 1987). Previously I showed that male guarding enhances hatching success. Unguarded eggs are lost because of predation by conspecifics and ants, and because they are washed off by rain, attacked by fungus, or disappear for reasons that have not been established (Mora 1990). Young Z. albomarainis undergo their first two molts in the nests (Mora 1987). Although males do not care for early instars, nests provide a secluded place for them before they disappear into the forest floor or the tree canopy.

I described nest construction behavior in a previous

study (Mora 1987, 1990). Besides building a nest, a male can acquire a nest by taking over an already occupied one or by utilizing abandoned nests. Nests are necessary for males to reproduce since females neither mate nor oviposit outside nests (Mora 1990). There are, however, nestless males in the population.

Females mature a few eggs at a time through the breeding season. When ready to lay eggs, they enter nests, court males, copulate, and then oviposit from 1-3 eggs inside the nest. They leave the eggs for the male to guard (Mora 1987). Females wander on tree trunks, and their behavior was poorly understood prior to this study.





14



The breeding season on Barro Colorado Island, where I conducted this study, lasts about six months from June to late December. This corresponds to the wet season on the island. The natural history of these animals during the rest of the year is unknown, but they probably live secretive lives on the forest floor, under crevices and on tree trunks until the rains begin. Laniatorid opilionids are known to live several years in captivity (personal observation) and I have observed marked individuals of this species reproduce in two consecutive years.


Study Site


I studied a natural population of harvestmen on Barro Colorado Island (BCI), Panama, from June to December of 1987 and 1988. I conducted a previous study of the same population from July to December, 1985 (Mora 1987). Barro Colorado is a biological reserve protected since 1923 and administered by the Smithsonian Institution since 1946. Its total area is 1500 ha; the island was isolated in 1914 after the Rio Chagres was dammed to form Gatdn Lake, which presently makes up the central portion of the Panama Canal.

The island is a hill rising 137 m above the Gatdn Lake. The hilltop is broad and flat, with steep (20-30 degrees) slopes at the edges. About half of the island is covered by young forest, a hundred or more years old, which is still growing back from old agricultural fields. Undisturbed, old





15



forest (200-400 years) covers most of the rest of the island (Figure 2-1). Both old and young forest have nearly 60 species of trees per hectare. On average 170 trees over 20 cm diameter at breast height (dbh) can be found per hectare. The height of the canopy in the old forest is 30-40 m (Foster and Brokaw 1982; Leigh 1982).

The most conspicuous seasonal changes in the island are associated with rainfall. Therefore, the biology of many organisms is closely tied to rains. The times when the wet and dry season start and stop are perhaps the most important aspects of the biological year on BCI (Rand and Rand 1982). These are independent of one another. The rainy season usually begins between mid-April to mid-May and lasts until December. Ninety percent of yearly rainfall occurs then (Dietrich et al., 1982). October and November are the months with the most hours of rain (Rand and Rand 1982). Rain falls usually in short but intense storms. The mean monthly precipitation during the rainy season is 310 mm, as compared to 51 mm during the dry season (data for 50 years, Dietrich et al., 1982). These wet-season rainstorms cause soils to swell, and in many places to become saturated and discharge water to small ravines scattered throughout the island (Dietrich at al., 1982). There are, however, dry periods of up to a month during the rainy season, sometimes causing the understory plants to wilt. This short dry season usually occurs during August. Over the past 50 years, the average annual rainfall was 2600 mm (SD=422 mm, Windsor 1990).





16



The mean monthly temperatures vary little annually.

Average annual temperature is 270 C in the open. Relative humidity remains high all year. The monthly relative humidity at midday, as measured by a sling psychometer, normally varies from 75-77% in March or April to 93% by November (Dietrich Bt al., 1982). Leigh (1982) described the climate, flora and fauna of the island.


Study Population


The population studied occupies the central plateau of the island. All trees containing nests have been mapped in an area of approximately 500 x 250 m. In 1987 I used three main study sites: (a) Site A consisted of a transect approximately 580 m long and 5 m wide, along short-cut trails at the central plateau of the island, (b) Site B consisted of a linear transect of approximately 400 m along Armour trail in the central plateau and (c) Log Site was a fallen log approximately 12 m long and 1.5 m wide, close to mark 15 on Wheeler trail. In 1988 I monitored Sites A and B and added a new plot, Site C, adjacent to Site A, measuring 500 x 154 m (Figure 2-1).




General Methodology


Most data I consider in the following chapters were

gathered from censuses of nests at the study sites. Trees





17



that contained nests were individually flagged. For all nests I recorded the height above the ground, the nest diameter (as the average of the maximum and minimum nest width) and the height of walls (the average of maximum and minimum wall height). I repeated these measurements whenever a new male was found occupying the nest. These measurements were taken following the protocol of Mora (1987).

Individuals were color-coded using various combinations of Testor's enamel paint on carapace and legs. Each individual was brought to the laboratory once and measured under a dissecting scope for body length (dorsally from occularium to posterior cephalothoracic spines), total width (at level of the fourth coxae), weight (to nearest mg) and total length of tibia + patella of the fourth leg.

Every other day all nests in all sites were surveyed,

recording the resident male as well as all females and other males present at nest sites, the number of eggs in the nest, number of juveniles, damage to nests, male and female behaviors, and interactions occurring at the time. All censuses were performed between 0800 and 1100 h. Additional data on nest visitation, courtship, and male-male and femalefemale interactions were recorded during about 200 h of opportunistic observations.

If a nest was abandoned for more than two weeks, only weekly checks were performed to confirm the status of that nest. After a take-over or a new occupation occurred, the new male was marked and measured. All females present in a





18



tree with a nest were monitored. Trees with no nests were not censused. I made opportunistic observations of females on trees that did not possess nests. I cannot sex subadult individuals; therefore, I neither marked nor measured them.




Definitions


Nest owner refers to the resident male of a nest and guards eggs inside it.

Nest is the physical structure built by males where eggs are oviposited.

Nest site refers to the tree where the nest is located. I expand the term in the last chapter to include an array of physical and biological variables. Resident female is a female who remains at a nest site for more than three consecutive days and assumed to have oviposited in the nest at least once. Residency reflects the outcome of female choice and was used as a measurement of nest and male success.

Study site is a plot delimited in the forest where I performed the censuses and most experiments (unless otherwise noted).

Male success is the breeding success of individuals, as measured by the number of eggs a male acquires, the number of eggs hatched at his nest and the number of resident females associated with the male during his nest tenure.





19



Nest success refers to the number of eggs acquired and hatched, and the number of resident females associated with the nest for the total time a nest is occupied during the breeding season.

Wandering female is a female not associated with a nest site. These were females found on trees with no nests. It also refers to females visiting nest sites and remaining there only 1-2 days.
























Figure 2-1. An estimation of the density of nests associated with forest composition on the Barro Colorado Island, Panama. The area monitored during 1987 and 1988 was located in the central plateau of the island and is outlined as the central, solid block. The location of the study sites A, B and C are shown in the lower left corner, with the Log Site indicated with an asterisk (*) Vegetation types were outlined after Foster and Brokaw (1982).








VEGETATION TYPES DENSITY OF NESTS old forest
abundant... young forest
2 sparce present clearings
3 rare or absent cut since 1914


141







BB /

:--;~ ;:: : : ::: :::~- .. ..... ............. : :: : h










kilometu-s














CHAPTER 3
NESTS AND MALE SUCCESS



The most remarkable aspect of the reproductive biology

of the tropical harvestman, Zvaopachvlus albomarginis is that males construct nests in which they guard eggs. Males build and repair nests with mud and tree bark, defend them, usurp them, and use abandoned nests (Mora 1987). The mating behavior of both males and females is strongly associated with nests: mature males and females spend the entire breeding season associated with them. I present here descriptions of the patterns of nest use and nest success and discuss how those patterns relate to the reproductive success of males. This identification of what contributes to the variation in male reproductive success is a first step toward understanding the evolution of male egg-guarding and the mating system of this species.

In this chapter, I will evaluate a male's reproductive success based on measurements of three components of his breeding success: male nest occupation, which measures the male's capacity to acquire and retain a nest; male egg acquisition, which measures a male's ability to attract mates, and egg hatching success, which measures the success of his paternal behavior. The number of females that associate with the males and their nests, and the length and


22





23



nature of these associations are also important components of the male's and the nest's success. These aspects are discussed in Chapter 4.

The difficulty of distinguishing male attributes from nest qualities when evaluating male success was a problem encountered throughout this study. The problem arises because females will not mate or oviposit outside nests. Hence, the success of a male is determined by his ability to acquire a nest, and the quality of the nest and nest site. At the end of this chapter I present results from an experiment that attempted to partition out the relative contribution to reproductive success of males and their nests.


Nests and Male Success: General Descriptions



Nest Distribution and the Sampled Population


Nests and individuals of Z. albomarainis were abundant

on Barro Colorado Island. I marked 223 nests and monitored a total of 158 nests. I followed 1031 individuals of the 1675 marked during the 1987 and 1988 breeding seasons (Tables 3-1; 3-2).

The distribution of nests on the island was not

continuous. Rather, nests were concentrated in the upper plateau where there was a combination of old and tall, young forest. Most areas with young forest, scrubby secondary





24



growth and those adjacent to the island edges had very few or no nests (Figure 2-1). There were additional areas on the island, of similar forest composition to the one of highest nest densities, that were not utilized as nesting areas. I consider all sites except the Log Site (a fallen log along a trail), as forest sites, with nests both close to and away from trails.

Nests were not more abundant on any particular tree

species. They were rarely found on small trees; only 34 out of 223 nests monitored were built on trees of less than 40 cm dbh (X2=107.7, p<0.001, df=l). These observations suggest that males established nests on mature trees, which, besides having a greater girth, perhaps provided greater microhabitat complexity by having more epiphytes, more food sources, and more complex architecture than young trees.

Contrary to previous observations (Mora 1987), pooled

data from 1987 and 1988 showed no difference in the number of nests found on buttressed and unbuttressed trees (X2=0.74, df=l, n=223, p>0.05). However, there were more aggregations of three or more nests on buttressed trees than on smooth ones (Table 3-3). This difference may be an artifact due to the fact that buttressed trees are generally larger in surface area than unbuttressed ones.

Based on my observations on the main causes of nest

destruction and egg mortality (Mora 1990, this study Chapter 6), some variables that may account for the differential use of nest sites are the amount of rainfall running down the





25



tree trunks, the probability of fungal infestations, and predator attacks at the nest site. Differences in canopy cover may produce differences in the amount of water running down the trunks. This flow of water not only destroyed eggs when it accumulated in the nest, but it also washed off the nest walls and nest floor. Nest sites may also differ in their probabilities of experiencing attack by fungus or egg predators. Males clean the nests and are able to deter the damage of fungus, but probably at a high cost (see Chapter 6).

The main egg predators are conspecifics (females and

males), several species of ants, and an unidentified species of flatworm. Several characteristics of the nest (height, orientation) may increase its likelihood of an attack and several characteristics of the sites (humidity, temperature, shade, forest composition) may attract some kinds of predators. The patterns of nest and egg damage due to these variables are discussed in Chapter 6. Measure of Success


The breeding success of an organism has various components. Among them, survival to breeding age, reproductive life span, fecundity, mating success and offspring survival have been proposed as important sources of variability in success among organisms of a given population (Clutton-Brock 1988). I cannot age individuals and have not





26



made observations between nesting seasons; my measures of success are derived from the observations and experiments on individuals present during the nesting season. This will lead to an overestimate of the average success of individuals in the population since it does not take into account the proportion of individuals that never reach reproductive age or fail to join the nesting population (Fincke 1988).

I measured components of male success related to their pre- and post-mating success. The pre-mating success of a male includes survival to breeding age, success in intrasexual competition, and mating success. Success in intrasexual competition refers to building, taking-over and finding abandoned nests, and at keeping a nest. Mating success is derived from the male's chances of receiving female visits and securing copulations at the nest. I estimate premating success from the male's ability to acquire and hold a nest and from the total number of females visiting and remaining at the site for more than three days. Male success ultimately derives from the patterns of female associations with the males and nests. These associations of females to males and nests and their relationship to male success are examined in Chapters 4 and 5.

The post-mating success of a male is defined as both

rate of egg acquisition and guarding abilities. It will be measured as both the number of eggs a male acquired during his residency at a nest and as the total number of eggs hatched from the nests. Both the pre- and the post-mating





27



success of a male are influenced by his expectation of further reproduction, and at any moment it is relative to the success of the other males in the population (Trivers 1972).

Since nests were used by different males, I have also investigated the success of the nests. This was done by determining the total number of eggs a nest acquired, the number of eggs hatched at the nest and the number of resident females associated with the nest. Nest success was estimated regardless of the identity or the number of nest owners.

In the following sections, I describe the patterns of nest use and nest success. Second, I describe the relevant natural history of males, their behavior, and patterns of success. Third, I present a male switching experiment that separates the male and nest effects on breeding success. All correlations presented are Pearson correlations, unless otherwise stated. Thus I am assuming both a linear relationship between the variables and the normality of the variables (Sokal and Rohlf 1979). The use of correlations for investigating the breeding success allows me only to establish associations between variables. They do not prove causation, which can only be established through experimental manipulations. The sample sizes for the tests reported will differ somewhat when considering different questions. This is because I do not have complete records on all aspects considered for all individuals and nests.





28




Patterns of Nest Use



Nest Re-use and Abandonment


Not all nesting males guarded eggs in nests that they built. Nests often remained intact and were re-used. Acquisition of occupied nests through take-overs, and using abandoned nests were ways of obtaining previously-constructed nests.

Use of nests from previous seasons. The permanence of nests throughout different breeding seasons was remarkable. Four out of the 172 nests monitored in 1985, were still present and were used in 1987. Three persisted until 1988. From Sites A and B, 21% of the nests (18 out of 86 nests marked in 1987) lasted one year.

All nests remaining from the 1985 season that survived to 1987 were re-used at least for short periods of time in 1988. A total of 12% of the nests marked in 1987 (Sites A+B=86 nests) were re-used in 1988. Of the nine nests from Site A (1987) still present in 1988, four were used throughout the complete season, and six of the nine nests from Site B (1987) were occupied throughout the 1988 season. A total of 56% of the nests from 1987 lasting to the next season were reutilized in 1988.

Take-over of nests. One way that males acquired an

already-constructed nest was by means of a take-over. Take-





29



overs involved fights between males (Mora 1987). I attributed new ownership as the result of a take-over whenever I found a new male at a nest without any prior period during which the nest had been empty. Thus the takeovers that I report here were the between-census replacements of one male by another. This criterion may overestimate the rate of take-overs in the population while probably underestimating the rate of nest abandonment.

Most of new male occupations at nests (63.5%, n=74, in 1987 and 56%, n=84, in 1988) were the result of nest takeovers (Table 3-4). The frequency of take-overs was not different in 1987 from that of 1988 (X2=0.39, d.f.=l, p>0.05; Table 3-4).

Nest abandonment. Forty six percent of all nests monitored in this study (n=158) were abandoned by their resident males. Some males abandoned nests after they suffered damage from rain, ants or fungus (43%, n=72 nests abandoned) or for unknown reasons (57%, n=72). Most males disappeared from nest sites (trees) after abandoning the nest (88%, n=72), but others remained in the same tree for the rest of the season (12%, n=72). A total of 37 (26.6%, n=158) new male occupations of nests in the 1987-1988 breeding seasons were on previously abandoned nests (Table 3-4). Therefore, about half of the nests that were abandoned (49%, n=72) were never used again that year.

Number of nest owners. Almost half the nests monitored in both seasons (44%, n=158) were occupied by different males





30



throughout the breeding season (Table 3-5). Multiple occupancy of nests was due to nest take-overs (36%, n=158) and use of previously abandoned nests (23%, n=158; Table 34).

Lenath of nest use. Nests were occupied by one or more males for an average of 2.5 months (Table 3-6). There was no difference in the length of nest occupation between years (t=-1.35, nl=59, n2=84, p=0.18, unpaired two-tailed t-Test).


Building New Nests


Nests were made out of mud and tree bark collected by

males from crevices of the trees and from deposits around the supporting structures of epiphytes. Nest construction behavior was previously described (Mora 1987). Males compacted the nest material by rolling it with their chelicerae and pedipalps into tiny balls that they applied directly to the tree. The floor was laid out first and consisted of a circular, single layer of material and then, walls were raised vertically around the circular base.

Even if a nest did not survive to the next breeding season, new males built nests at the same locations year after year. This occurred for four nests from the 1985 plot, three nests in the B Site, and two in the A Site. Four percent of the nests monitored during the two seasons (n=158) were rebuilt by their occupants at distances of 7-16 cm from their previous nest. At the forest sites, 82% of the nests





31



marked in 1988 were new (80 out of 98). Considering the availability of trees in this population, site fidelity was remarkable. These observations suggest that some trees, and particular locations on trees consistently attract reproductive males.


Nest Measurements


Nests had a mean height from the ground of 50.3 cm

(SE=4.9, range 8-171 cm, n=98, Table 3-7). They had a mean diameter of 3.2 cm (SE=0.13, range 2-7.8, n=152, Table 3-7) and their walls had a mean height of 0.5 cm (SE=0.24, range

0.14-1.1, n=152, Table 3-7). These dimensions were not different from those of nests sampled in 1985 (Mora 1987). Nest height above ground was not correlated with either nest diameter (r=-0.07, n=98, p=0.52) or the height of the walls (r=0.14, n=98, p=0.23). Also, the diameter of the nest was not correlated with the height of the walls (r=0.21, n=152, p=0.08). There were no differences between the two seasons in the height at which males built nests (Table 3-7). There were differences between years in the sizes of the nests: males constructed nests of greater diameter in 1987 than in 1988 (Table 3-7), and nests had higher walls in 1988 than in 1987 (Table 3-7). These differences were probably an artifact of the differences in quality of the calipers used to measure nests in the two breeding seasons.





32




Patterns of Nest Success



Is Eag Laying Random?


Although some nests appeared to receive more eggs than others, this did not necessarily imply that oviposition was biased towards some nests. Were there some nests more likely to acquire eggs?

Methods. To evaluate this question, I compared the distribution of ovipositions for a large sample of nests throughout a discrete period of time to that expected if the ovipositions were random. If the distribution of oviposition events was not random, more nests than expected would have no oviposition events and fewer visits (and then, "bad" nests do indeed occur), and more nests than expected would have a larger number of oviposition visits (and thus significantly better than the unsuccessful ones). There are several known distributions to which we can compare the observed distribution of ovipositions. Because many nests experience few or no visits (oviposition events were rare), the Poisson distribution was appropriate (Sokal and Rohlf 1979).

I selected nests monitored in 1988 and evaluated the number of ovipositions experienced by 50 nests over a 30 d period. I analyzed data from September because that month had a large number of resident males present at nests for the 30 d period. I calculated the number of ovipositions as the





33



ratio of the number of new eggs over the mean number of eggs per oviposition (x=3, Mora 1987). The distribution of ovipositions was compared to an expected Poisson distribution by a two-tailed Chi-squared test.

Results. The distribution of oviposition events per nest was significantly different from the expected Poisson distribution (x=5.06, SE=0.78, X2=64.2, d.f.=3, p=0.001, Figure 3-1).

Discussion. Oviposition events were clumped. There

were more nests with larger numbers of eggs than predicted by the Poisson distribution. There were some nests/males more successful at getting ovipositions; therefore, "good" and "bad" nests may in fact exist. However, the distribution of ovipositions is not bimodal. What may be the variables accounting for the differences among nests? What are the patterns of nest success in this population? These questions are considered in the following sections. Were Old Nests More Successful Than New Ones?


Success in 1987. The nests that survived and were used in 1988 acquired more eggs than average in 1987 (for Sites A and B, t=3.67, nl=18 nests that persisted to the 1988 season, n2=20 nests that did not persisted, p=0.001, unpaired, twotailed t-Test). However, nests that persisted and were used in 1988 showed no difference in hatching success in 1987 as





34



compared with the nests that did not persist (t=0.74, nl=18, n2=20, p=0.46, unpaired, two-tailed t-Test).

Success in 1988 Old nests (18 nests from 1987 that survived and were reused in 1988 in Sites A and B) accumulated a mean of 29.6 eggs (SE=5.5, n=18) and hatched a mean of 61.4 juveniles (SE= 15.9, n=18). The number of eggs accumulated at these old nests was not different from the mean number of eggs at the new nests sampled that year in the C Site (Table 3-3, t=0.84, nl=18, n2=76, p=0.41, unpaired, two-tailed t-Test). However, old nests in 1988 produced significantly more juveniles than new ones (mean juveniles for new nests= 27.7, t=2.11, nl=18, n2=76, p=0.05, unpaired, two-tailed t-Test).

The success of a nest in a year did not predict its

success the following year. The number of eggs accumulated in the 18 nests monitored in 1987 that survived and were reused in 1988 was not correlated for the two years (r=0.16, n=18, p=0.6). Nor did the number of juveniles produced in 1987 correlate with those produced in 1988 in the same nests (r=0.4, n=18, p=0.18). Another measure of success, the association of females to those nests, will be discussed in Chapter 4.

In summary, nests from the 1987 season that persisted and were reutilized in the 1988 season, had been successful nests in 1987 in acquiring eggs. Those same nests in the following year showed a higher hatching success than new nests. The analyses showed that the success of any one





35



individual nest in one year was not correlated with its success in the following season. Whether or not nests are consistently successful from year to year is not clear. From two seasons of observations, the pattern is that nests that are surviving and re-used were more successful than average (in terms of egg accumulation or hatching success) during both nesting seasons. However, the success of these nests was not consistently accounted for by either a higher number of eggs accumulated or by a higher hatching success.


Was Lenath of Occupancy Associated with Nest Success?


In both years, the length of nest use within a season was correlated with both the number of eggs accumulated at the nests and the number of juveniles hatched in them (Table 3-8). These positive associations may have come about because more successful nests were occupied longer or because males' longer occupancy of nests increased the success of the nests. I have no way to distinguish between those two explanations.


Are Take-over Nests More Successful?


Nests that experienced take-overs held higher numbers of eggs (mean for nests that were taken over=37.9 eggs, SE=3.2, t=5.06, n=53, p=0.001, two-tailed t-Test, compare to averages in Table 3-12) and hatched higher numbers of juveniles than average nests in both seasons (mean for nests taken-





36



over=39.5, SE=6.6, t=3.2, n=53, p=0.002, compare to averages in Table 3-12). These data suggest that male competition for nests, estimated by the rate of take-overs, was higher at more successful nests.


Is Nest Success Correlated with the Number of Nest Owners?


General correlations. There were no correlations between the number of owners a nest had throughout the breeding season and either the number of eggs that accumulated at the nest (r=0.08, n=101, p=0.41), or the number of juveniles produced at the nest (r=0.001, n=101, p=0.99). Since most new owners were due to take-overs (Table 3-4), this may seem contradictory to the observations indicating that nests experiencing take-overs were more successful than the average nests. Two reasons may account for the lack of correlations between the number of male residents and the success of the nests: (a) a male eats the eggs present at the nest after a take-over, decreasing the counts of eggs and juveniles accumulated at the nests and (b) new ownership of a nest may occur after a nest had been abandoned for some time. Therefore, many of the multiple occupancies occurred in nests that were empty for prolonged periods and thus had a lower accumulated number of eggs and juveniles for the season.

Success of nests with more than three owners. Most

nests were occupied by one or two different males, and a few





37



had three or more resident males through the season. A comparison of the number of eggs and the number of juveniles between the 14 nests of this study that had three or more owners (Table 3-5) and a random sample of 14 nests with one or two owners (generated with a table of random numbers) showed the former had a significantly higher number of eggs (Mann-Whitney U Test, U=69, U crit=64, p<0.05), and hatched a higher number of juveniles (U=64, U crit=64, p<0.05; Figure 3-2).

These results indicate that males were indeed reusing

more successful nests. As a consequence, the higher success of nests with more owners may be simply that there were more breeding attempts at those nests. I cannot distinguish if nests were more successful because they were reutilized more times and accumulated more eggs and juveniles, or if they were reused more times because they were indeed better nests than average.


Are Abandoned Nests Less Successful?


The success of abandoned nests was not different from the success of the average nest in the population. I have excluded the data from 1987 from this analysis because I do not have complete records for all nests abandoned in 1987. Abandoned nests in 1988 neither accumulated fewer eggs than the population mean (mean for abandoned nests=28.8, SE=4.2, t=1.77, n= 17, p=0.09, compared to the population mean in





38



Table 3-12) nor hatched fewer juveniles (mean for abandoned nests=24.4, SE=6.1, t=0.97, n=17, p=0.34, compared to population mean in Table 3-12) than continuously occupied nests. These observations suggest that males were not abandoning the less successful nests in the population.


Is Nest Success Correlated with Nest Measurements?


In general, the success of the nests was not strongly associated with most of the measured nest dimensions (Table 3-9). There was a weak correlation between the diameter of the nests and the number of eggs that nests accumulated, suggesting that nests of larger diameter may hold more eggs than smaller nests.

The only nest measure associated with the success of the nests was the height of the nest's walls. Nests with higher walls hatched more juveniles than nests with lower walls (Table 3-9). There are several reasons why walls may enhance the hatching success at the nests. Walls are physical obstacles that may deter the attack of predators such as ants and flatworms. Walls may also block water running down the tree after a storm and prevent eggs from being washed away. Such excess moisture might also increase fungal growth among the eggs.





39




Male Life History and Behavior: General Descriptions



Male Size


Males had a mean weight of 0.014 g (SE=0.0002, range

0.008-0.2 g, n=227). Their mean total body length, measured from ocularium to the posterior cephalothoracic spines, was 3.02 mm (SE=0.04, range 2.24-4.0 mm, n=227) and mean total body width, measured at the level of the fourth coxae was 2.5 mm (SE=0.03, range 1.6-3.12 mm, n=227). The mean length of the fourth tibia+patella was 3.67 mm (SE=0.01, range 3-4 mm, n=227). This measurement most accurately describes the body size in opilionids (M. Goodnight, personal communication) and varied little in this population. Male Longevity


Adult males can survive several years and reproduce for at least two seasons. From the 156 males marked at the A and B sites during the breeding season of 1987, four were recovered in 1988. Three of these had been occupying nests since the beginning of the season. The real proportion of individuals surviving to a second breeding season may be higher than this since the enamel I used for marking animals is not permanent. Individuals that survived to a second year can potentially reproduce throughout the complete breeding season. One take-over that occurred in November of 1988 was





40



accomplished by a male marked mid-season in 1987, and one nest owner in his second year held a nest for the complete 1988 season.


Length of Male Residency at Nests


The average residency time of a male as a nest owner was 48 d for all sites over both seasons (Table 3-10). This is only about a fourth of the total nesting season (JuneDecember). Both the shortest and the longest residence times recorded for males were in nests at the Log site where there was a high concentration of nests on a single tree. This may have generated intense competition for nests and produced both very short and very long nest-tenures.



Origin of New Residents


Most (67%, n=94, Table 3-4) of the new owners of nests were not prior residents on the tree where the nests were located nor were they from trees adjacent to those nests. Rather, they were mostly unmarked males new to the nest site (Table 3-11 , X2=10.88, df=l, p< 0.001). It was easy to find males since I sampled the nest sites carefully, including the tree crevices, up to a height of 2 m. There is a possibility that the new nest owners had come from the tree canopy, but individuals in this population were not commonly found at heights over 2 m (Mora 1987).





41



I suggested previously (Mora 1987) that non-guarding males stayed at nest sites probably waiting for an opportunity to take-over the nest or to use it after it was abandoned. Accordingly, I had expected that most of the new nest owners would be males marked at, or close to, the nest site. My observations however, showed that those males that had associated with the nest sites prior to the new occupancy were not necessarily the ones replacing the nest owner. Male movements were rather extensive, since the majority of new nest owners found between censuses were males not seen previously in the nest area.


Nestless Males


It is most intriguing that not all males built nests in the field. Most males built nests in captivity when a piece of bark was provided. I cannot estimate the proportion of nestless males since my observations were limited to nesting sites, but my impression is that it is large. I speculated (Mora 1987) that salivary secretions, which are added to the nest material at the moment of applying it to the tree (one can see the strings of silky material with the aid of a flashlight), may be metabolically costly or may develop only at a certain age. If the ability of males to build nests is constrained, this might account for the population of nestless males.





42



In 1987, my assistant and I searched for the sources of nest building secretions by performing dissections of the cephalothoracic region of males that had finished constructing nests (n=4), of males collected at the moment of building a nest (n=2), and of wandering males not associated with nests (n=4). We found no evidence of a specialized gland. However, we did not utilize specific dyes, fixatives or equipment that may have enhanced our possibilities of finding such a gland, so its existence cannot be discounted. If they exist, the glands are probably very small and very fragile. Besides the repugnatorial glands located at the base of the first and second coxae, there are no accounts of accessory glands in opilionids (Berland 1949; Hillyard and Sankey 1989).


Male Behavior: Backaround


The range of behavior patterns shown by z2. albomarainis is remarkable for an arachnid. The following account is of the activities of nesting males only. I have indicated with an asterisk (*) the behavior patterns for which descriptions were provided in Mora (1987). Each one of these behavioral patterns is a source of variability in breeding success of males. The relationship of some of them with male success will be evaluated in a following section.

The different kinds of behavior patterns performed by males can be grouped into four general categories. (a)





43



Foraging behavior. Males leave the nests to forage for food and nest materials. Males are omnivorous; they feed on insect larvae deposited in crevices, on carcasses of dead insects, on fruits and they hunt termites. They also leave the nests to gather mud and tree bark that they utilize for nest construction and repair. (b) Nest acquisition. Includes nest construction(*), take-overs(*) and searching for abandoned nests. The mechanisms and the mode of searching for nests and nest sites are not known. (c) Courtship and copulation. Male mate choice is important in this system since not all female courtship attempts lead to copulation and males chase off the nest some females that have initiated courtship (Chapter 4). (d) Eaa guarding and nest maintenance. Guarding activities of males include nest cleaning(*), chasing off predators(*), and repairing nest damage(*).


Components of Male Success



Ega Acauisition


Males guarded a mean of 21.4 eggs (SE=1.3; range 1-102, n=227; Table 3-12) during their residency time. Nesting periods, however, were not equal for all males. A better way to compare the success of males takes into account the residency times of males at nests. Correcting for the males' nest tenure, males acquired only one new egg every 2 d (x=0.5





44



eggs/d, SE=0.06, n=227; Table 3-12). Both measurements of success underestimate the number of eggs acquired by males since many eggs are lost to predation (Mora 1987). There was no difference in the number of eggs acquired by males over their residency time between seasons (t=1.87, nl=119, n2=108, p=0.06, two-tailed unpaired t-Test, Table 3-12). Hatching Success


Nesting males hatched a mean of 18.5 juveniles over their residency time at nests (SE=2.3; range 0-211, n=227; Table 3-12). Taking into account the male residency time at nests and considering the total juvenile production per day, males hatched, on average, 0.3 eggs/day (Table 3-12).

The 1988 season yielded higher nest production of

juveniles than did the 1987 season (t=5.51, 1987=119 nests, 1988=108 nests, p<0.001, unpaired, two-tailed t-Test, Table 3-12). The differences between sites in number of juveniles hatched were not significant (Table 3-12) except for the comparisons between the A and B Sites in 1987 (t=2.06, n=55, p=0.04, unpaired, two-tailed t-Test). Relationship Between Eaas Acauired and Egs Hatched


The mean number of eggs at nests showed a positive

correlation with the number of juveniles produced (r=0.47, n=201, p<0.001). There was a stronger correlation during the 1988 season (r=0.52, n=112, p<0.001) than for the 1987 season





45



(r=0.46, n=89, p<0.001). The more eggs a male guarded during his tenure at a nest, the more juveniles hatched from the nest (Table 3-4, Figure 3-3). Some males appeared to have hatched more juveniles than the mean number of eggs they guarded (Figure 3-3). This result was due to the great variability of the data. Many eggs never hatched or disappeared from one census day to another. Because of those fluctuations in egg numbers at nests from census to census, I estimated egg acquisition from the mean number of eggs a male guarded during his residency at a nest. Since guarding males will only care for their own eggs (Mora 1987), juveniles hatched were most probably the males' offspring.


Success and Male Size


There were no significant correlations between the success of the males and any of their phenotypic characteristics. There was however, a tendency for large males to be slightly more successful than small males. There were overall positive but insignificant associations for both years between male weight and the number of eggs acquired (r=0.21, n=143, p=0.09), male weight and the number of juveniles produced (r=0.26, n=143, p=0.27) and the length of tibia+patella 4th and the number of juveniles hatched at the nest (r=0.36, n=143, p=0.07). Despite the fact that none of these associations was significant at the 0.05 level, they





46



suggest that male size may have some influence on the success of nesting males.

These are not surprising results. Body size has been recognized as contributing to male breeding success in many species (Clutton-Brock 1988). In this species, large body size may confer a male advantage in competition for nests and in preventing predator attacks. Success and Male Longevity


The success of the four males marked in 1987 that

reproduced again in 1988 was not different from the average success of males in those years. Those four "old" males acquired the same number of eggs and hatched the same number of juveniles as average males in 1987 (for eggs: t=0.04, n=4, p=0.96; for juveniles: t=-0.1, p=0.92, one sample t-Test) and in 1988 (for eggs: t=0.97, n=4, p=0.4; for juveniles: t=2.04, p=0.13, one sample t-Test).

Males reproducing for a second season had higher success in the second than in the first year. The four males followed for two consecutive seasons guarded more eggs (z=1.6, p=0.05, one-tailed Wilcoxon Matched-pairs Signed-rank Test) and produced more juveniles (z=-1.83, p=0.03, onetailed Wilcoxon Matched-pairs Signed-rank Test) in 1988 than in 1987. This pattern of success is not uncommon. It has been recognized that several components of reproductive success improve with age in Drosophila flies (Partridge





47



1988), frogs (Howard 1988) and great tits (McCleery and Perrins 1988) among others. Since harvestmen do not grow after their last molt, part of this improved success may be derived from increased parental experience as occurs with great tits (McCleery and Perrins 1988), or to their improved abilities to recognize and hold a good nest. Survival itself could be considered proof of a vigorous constitution, which may also result in good parenting and female preference for those males.


Success and Male Residency Time


The length of nest tenure influenced the success of

males. Although the length of nest use was correlated with the number of eggs accumulated in the nest (Table 3-8), there was no correlation between the length of male residency at the nest and the mean number of eggs that the males acquired (r=0.13, n=101, p=0.19) during their nest tenure.

Residency time was significantly correlated with the

number of juveniles hatched (r=0.54, n=101, p< 0.001). These results suggest that a long residency time may not increase a male's chance to accumulate eggs, but it increases the number of juveniles hatched. Longer residencies may decrease the levels of predation by conspecific females. This could confer an advantage to long-term residents and may select for males holding onto nests for as long as they can. The intriguing question of why males abandon nests despite this





48



strong association between residency time and hatching success will be addressed below. Success and Male Behavior


It is possible that the guarding abilities or the mating success of males did not depend on size, but rather on their behavior at the nests. Are there any differences in nestrelated behaviors of successful as compared to less successful males?

Methods. In 1987 season, my field assistant and I made continuous focal male observations at the Log site. I categorized males as "successful" when they had a hatching success of more than 20% (mean+ SD for all males at the Log Site) and at least one resident female. Males in nests with lower hatching success and no resident females were categorized as "unsuccessful". This categorization was done prior to the observations. We performed paired, simultaneous 3-h block observations on eight males, four in each category. The pairs were chosen using a random number table. The time of day for observation was randomized. Each nest was observed twice, for a total of 1440 min.

Results. Only successful males repaired walls even

though both successful and unsuccessful males brought mud and bark to the nests (Table 3-14). Copulations were the only other activity exclusive to successful males. There was no difference in the number of times that successful and





49



unsuccessful males engaged in nest cleaning activities (MannWhitney U Test for the amount of time spent cleaning, U=5.2, U crit.=16, p> 0.05), nor was there a difference in the amount of time males spent out of the nest (Mann-Whitney U Test, U=5.7, U crit=16, p>0.05, Table 3-14). Only unsuccessful males were observed chasing off other individuals.

Discussion. Aggressive interaction was observed only in an unsuccessful male. This challenges my hypothesis that successful nests were competed for more intensely (as reflected by the higher rate of take-overs, which involve aggressive interactions) than unsuccessful ones. Aggressive interactions may diminish at nests after long residency or they may be too infrequent to be recorded in the time assigned for these observations.

There were only two behaviors that distinguished successful from unsuccessful males: the occurrence of copulations and nest wall repair. I have shown before that the height of nest walls was the only nest characteristic associated with hatching success at the nests. It is not surprising then, to find that males that repaired walls more often were more successful than males that did not. I would predict that females may cue in on the condition of nest walls when assessing males. Manipulation of the height of the walls and its effect on the reproductive success of males could be an area of interest for future studies.





50




Why do Males Abandon Nests?


The possession of a nest is necessary for males to

reproduce. Among the lines of evidence that suggest that there is competition for nests among males are the high rates of take-overs, the occurrence of nest reuse (Table 3-4) and the difference in success among nests (Tables 3-2, 3-9). I have also shown that the length of nest tenure is associated with hatching success at the nest. If nests are indeed difficult to acquire (there are nestless males in the population and there are take-overs and nest re-occupation after abandonment) and if longer tenure increases the hatching success, the question arises, why do males abandon nests?

It is expected that the decision of whether to abandon a nest will be determined by the benefit of remaining at the nest as opposed to what a male will gain by leaving. The data one needs to evaluate male decisions on whether or not to abandon a nest are the fitness outcomes for (a) remaining in the nest and (b) alternative behaviors. We need to know the probabilities of a male acquiring another nest, the probabilities that the new nest will be better than the one he already has, the costs for searching for or taking-over a nest, the costs of keeping a nest for extended periods, the remaining life expectancy for these males, and how all the above change with the age of the individuals and throughout





51



the breeding season. For this kind of analysis, I only have data on the take-over rate of nests.

Another way to understand why males abandoned a nest

site is to evaluate whether the males' probability of getting eggs changes over time. By doing this I am reversing the question and am now asking why males stay at nests. If the probability of acquiring eggs decreases with time we would expect males that have not acquired eggs to leave the nests. Conversely, if the probability of getting eggs increases with time, we expect males to keep nests for as long as they can. If the probability does not change with time, we expect males not to base their decisions of staying or not on the daily nest success.

I have constructed a decay curve for the time it took a sample of males to get their first egg. The curve plots the log of the number of males without eggs versus time, following a group of males over 30 d and resembles the survivorship curves familiar to population ecologists. I have drawn curves for early-, mid- and late-season 30 d periods in 1988 (Figure 3-4). These curves showed a slow and relatively constant decay in the number of males with no eggs. The shape of the curves was the same throughout the season. They mean that there were about equal probabilities that each male will get eggs each day and that those probabilities did not change through the season. Consequently, the number of days a male remained without eggs should not influence his decision of whether or not to leave





52



a nest. These results allow me to explain why males did not leave unsuccessful nests, but the question of why they did leave successful nests remains unanswered.

One important reason that males left nests was the

occurrence of nest damage. Nest damage was due mainly to ant and fungus attacks and heavy storms. The incidence of such damage and their possible effect on males' decisions to abandon nests will be discussed in Chapter 6.






Nests or Males? A Male Switching Experiment


I have been interested not only in identifying the

extent of the variation in male reproductive success, but also the phenotypic and ecological variables that account for the observed differences. Such information is necessary to evaluate the operation of natural and sexual selection and the evolution of morphological and behavioral adaptations. In this species, the main problem with addressing the question of what variables may be important for explaining variation in male success is separating nest and male effects since male success is dependent on the possession and retention of a nest. Addressing this problem is also important because nest success also reflects the reproductive success of the females that oviposit in them (Chapter 4).





53



I attempted to separate the effect of the nest from the success of males by performing a male switching experiment. I switched both successful and unsuccessful males to previously successful and unsuccessful sites and compared their success, in terms of the number of resident females retained, before and after the switching.

If male success is mainly determined by the nest, the success of an unsuccessful male will increase when switched to a successful nest. If females are associated with the nest rather than with the males, they would not leave and would continue ovipositing in a nest regardless of what male was present. The specific predictions were that if the success of males prior to the switching was largely due to the nest, it would decrease when switched from a good to a bad nest. If the success of males was not strongly associated with the nests, the success of males would not, on average, change when switched to a new nest.

Methods. I considered successful males as those with more than one resident female (a female associated with the nest for more than 9 d, i.a., that will potentially oviposit twice), and with a hatching success 20% or more, i.j., above the mean for the population at mid-breeding season. Bad and good nests were defined using the same criteria. In September of 1988 (mid-breeding season), considering nest sites with only one nest per tree, I was able to identify a total of 15 successful males at good sites and 18 unsuccessful males at bad sites. I randomly assigned males





54



to new locations. Seven successful males were switched to good sites and eight were switched to bad sites. Of the 18 unsuccessful sites and males identified, eight males were switched to previously successful sites and 10 were switched to unsuccessful sites. Only one pair of males was exchanged through this random assignment. By randomization, one individual was to remain in its original site (one time in this experiment). I gently picked it up and removed it for 4 h before returning it to its original nest. I measured the success of males before and after the switching by the number of females that remained at the nest sites because males fed on the eggs present at the nest after their arrival at a new nest, and because it takes a variable and long time for the males to accumulate and hatch eggs in a new nest. Data on the number of females resident on trees before and after the switchings for each nest, were compared using Wilcoxon Signed-Rank Test, one-tailed, at an alpha level of 0.05.

Results. When males were switched, successful males decreased their success, in terms of number of females potentially ovipositing at the nests, when switched to bad sites (Wilcoxon Signed-Rank Test, t=0, n=10, p<0.001, onetailed). Conversely, previously unsuccessful males experienced an increase in the number of females potentially ovipositing at their nests when switched to good sites (Wilcoxon Signed-Rank Test, t=0, n=8, p<0.05, one-tailed, Figure 3-5).





55



Discussion. These results suggest that the success of

males depends on the nest/nest site. A male can increase his success by associating with a good nest. The quality of nests may be the most important factor accounting for the variance in male reproductive success.






Discussion: Patterns of Nest and Male Success


I have utilized three measures of success to discuss the patterns of nests and males success. These measures are the male occupation of nests (reflected as length of nest use when discussing nests or tenure of nests when discussing males), the number of eggs acquired and the total number of juveniles produced. Additional measures of nest and male success can be derived from the association of females to nests/males and will be discussed in the next chapter. The most important pattern emerging was that nest site was the most important component of the success of nests and males.

There are four general assumptions implicit in my

evaluation of the variation in male success. (a) A resident male fertilized all the eggs he was guarding. This assumption is supported by my observations that males fed on eggs that were not their own and they invariable copulate prior to oviposition (Mora 1987). (b) Eggs accumulated at nests were oviposited by the resident females. The number of





56



females visiting nests between censuses is unknown and their contribution to nests is not considered in this study. The validity of this assumption is evaluated in Chapter 4. (c) Males with no nests had no success. I have not observed females mating or ovipositing outside a nest. (d) Individuals and nests marked and followed were a random sample of the total population of nesting males.

Patterns of nest use. The distribution of nests on the island was patchy. Not only were nests concentrated in the central portion of the island (Figure 2-1), but within the study areas, males built nests in the same locations or a few centimeters away from former nests year after year. There was a high rate of nest reuse; nest take-overs and the utilization of abandoned nests resulted in the multiple occupancy of nests. Half of the nests monitored were abandoned at least once during a breeding season and more than half of the nests experienced at least one take-over (Table 3-4). There were practically no nests that stayed the same for a complete season. These observations also suggest that there is something about a nest or the nesting site that attracts males and results in the reuse of some nests.

Patterns of nest success. The number of ovipositions

among nests was not random. More nests received very few and more nests received more visits than expected if ovipositions were random (Figure 3-1). This suggests that there were indeed "bad" and "good" nests in the population.





57



Nest longevity was a source of variation in nest

success. Nests from the previous year (marked in 1987) were more successful in 1988 than nests constructed that year. The length of nest occupation also produced differences in nest success. The longer a nest was occupied, the more juveniles hatched in the nest (Table 3-8). Two observations suggest that "good" nests experienced higher levels of competition among males: nests that suffered take-overs were more successful than nests that did not, and nests that experienced multiple occupancy, i.e., with three or more nest owners, were more successful than nests that were not taken over or had only one or two different male residents (Figure 3-2).

Patterns of male behavior and success. The variation in male success was high when measured both as the number of eggs males guarded (Table 3-12) and as the number of juveniles hatched at the nests (Table 3-12).

Males traveled great distances searching for nests; most new owners were new to the nesting site (Table 3-11). Surprisingly, I found few nest-related behavioral differences among successful and unsuccessful males. Apart from copulations, the single behavior at the nest observed only in successful males was repairing nest walls (Table 3-14). Since height of nest walls was the only characteristic of the nests correlated with nest success (Table 3-9), the amount of time and the frequency that males repaired walls may influence male success. A male's ability to repair walls may





58



be related to his age or breeding experience. Repairing a nest wall implies the same process as constructing a nest; both require the addition of salivary secretions to compact and cement the nest material. The availability of such secretions may be functions of age or physiological condition. Dissections of males did not disprove that there may be a constraint on nest construction. Reutilization of nests suggests that nest construction may be costly for males. The association between height of walls and nest success may be due to higher walls preventing egg loss, from females choosing to oviposit in nests with high walls because they reflect parental abilities or the genetic quality of the male, or from both. The ability of a male to maintain the nest walls may also be affected by the quality of the nesting site. Sites likely to receive large amounts of rain running down the tree trunks or exposed to the attack of predators may be harder to maintain than safe sites.

It was surprising to find no strong association between size and success of males. Body size has been identified as an important component of the male's reproductive success in most species as larger size confers competitive advantage to males (Clutton-Brock 1988). Males of Z. albomarainis do not follow this pattern. This is not surprising for an arthropod. Males of most species of insects are smaller than females and rely on other visual, acoustic or chemical signals to compete with each other. Also, in species with high paternal investment other phenotypic and ecological





59



attributes affect breeding success. In many birds in which the male contributes to the caring of the young, nest site quality and breeding experience are the two main factors affecting the male's breeding success and not body size (Arak 1983).

Males can survive through and reproduce for at least two seasons. This means that older, more experienced males are part of the breeding population, together with young males in their first breeding attempts. Males that reproduced a second year had higher success than the average males in that year. Although the sample size was small and there was no evidence of behavioral differences among old and young males, these observations suggest that age was an important factor affecting the breeding success of males in this species. One possible manner in which age is related to male success is that old males are experienced and have proven survival attributes. The same physical attributes that permitted longevity may be important for increased reproductive success.

Is is unclear why some adult males in the population do not build their own nests. The size of the population of nestless males is unknown but is presumed to be extensive; they apparently have no reproductive success. The males' tenure at the nest site was associated with their success; males that stayed the longest had higher success (or vice versa). Since the probability of a male getting his first egg at a given site does not change with time during the





60



season (Figure 3-4), males should be selected to keep a nest for as long as they can. Why males abandon nests is not known. The success of the nests abandoned was not different from the average nests in the population. The extent to which nest damage may have caused nest abandonment will be discussed in Chapter 6.

The male or the nest? There are several reasons that

suggest the nest(site) is an important factor affecting male success. The high rate of nest re-utilization and the fact that not all males built a nest suggests that nests are expensive for males. The use of the same spots for two seasons suggests that not all sites are suitable nesting sites within the study area. The male switching experiment showed that females stay with nests after a male switching. Thus, the success of males was directly affected by the quality of the nest/nesting site.

A clear pattern emerges that variation in quality of nests and nesting sites may be the most important factor contributing to the variance in male reproductive success. The strong competition for particular nest sites, evidenced, among other ways, in the high turnover of nest owners may suggest that sites themselves are important in male decisions to challenge an owner and female decisions to stay at a particular nest site. The importance of the site for a female's decision to remain associated with a particular nest will be discussed in Chapter 5; the factors that may account for differences among sites will be discussed in Chapter 6.






61







Table 3-1. The number of nests marked and monitored for the
1987 and 1988 breeding seasons.


No. of nests Total No. Year Site monitored nests marked
1987 Log 36 39
A 22 56 B 16 30
Totals 74 125


1988 A 11 11
B 15 15 C 58 72
Totals 84 98


Overall Totals 158 223






62







Table 3-2. The number of individuals marked and followed, by site and sex for the 1987 and 1988 breeding seasons.


Individuals Followed Individuals Marked Year Site Males Females Males Females
1987 Log 63 102 79 >220*
A 40 236 57 361 B 16 38 20 45 Totals 119 376 156 >626


1988 A 12 36 16 66
B 20 73 38 120 C 76 319 183 470 Totals 108 428 237 656


Overall Totals 227 804 393 1282
* I marked approximately 40 more females at this site.





63








Table 3-3. Nest densities on buttressed and unbuttressed
trees for the 1987 and 1988 field seasons combined.


Number of nests on tree
Tree Trunk Type 1 2 >3
Buttressed 60 11 34 Unbuttressed 109 7 2


X2= 42.9, df=2, p<0.001






64







Table 3-4. Nest re-use for the 1987 and 1988 breeding seasons.

Mode of New Utilization
Used after
Year Site Take-over Abandonment n
1987 Log 11 17 36
A 16 3 22 B 0 0 16 Total 27 20 74


1988 A 4 4 11
B 3 2 15 C 23 11 58 Total 30 17 84


Overall Totals 57 37 158






65







Table 3-5. Number of males occupying the nests monitored
during the 1987 and 1988 breeding seasons.



Number of Males Occupying Nest Year Site n 1 2 3 >4

1987 Log 36 20 11 3 2
A 22 7 12 2 1 B 16 16 0 0 0 Totals 74 43 23 5 3


1988 A 11 7 2 1 1
B 15 10 5 0 0 C 58 28 26 4 0 Totals 84 45 33 5 1


Overall Totals 158 88 56 10 4






66






Table 3-6. Total time (in days) nests were occupied by one or
more males during the 1987 and 1988 breeding seasons. Year Length of nest use-SE (d) range n

1987 82.4 � 5.27 8-161 59 1988 75.8 � 4.85 6-160 84 Overall 78.7 - 3.57 6-161 143














Table 3-7. Measurements (x � SE, in cm) of nests monitored during the breeding seasons of 1987 and 1988. t values and probabilities refer to comparisons between years using unpaired t-Tests, two-tailed.



Year Nest Measurement 7 - SE (cm) t p n 1987 Height from ground 48.1 � 7.0 -0.4 0.71 34 1988 51.9 � 6.8 64

1987 Diameter 3.5 � 0.4 3.35 0.001 68 1988 2.9 � 0.1 84

1987 Height of walls 0.2 � 0.4 -4.3 <0.001 68 1988 0.6 0.03 84





68







Table 3-8. The correlation between the length of nest use (d)
and the number of eggs accumulated at the nests and the
number of juveniles hatched (Pearson correlation coefficients).


Year Variable r n p
1987 Eggs 0.35 59 0.008
Juv. 0.39 59 0.004

1988 Eggs 0.29 84 0.01
Juv. 0.66 84 <0.001

Overall
Eggs 0.29 143 0.006 Juv. 0.45 143 0.001






69







Table 3-9. The correlation between the nest measurements and
the number of eggs accumulated in the nest throughout the season (Eggs) and the number of eggs hatched (Juv.) at the
nests. (Pearson correlation coefficients).



Nest Variable n Eas p Juv. 0
Height from
ground 98 0.07 0.66 0.12 0.48

Diameter 152 0.30 0.06 0.07 0.68

Height of
walls 152 0.10 0.50 0.35 0.02






70








Table 3-10. Mean residency times (d) of nest owners for the
1987 and 1988 breeding seasons.


Year Site Mean davs(SE) ranae n
1987 Log 49.1� 5.03 3-150 63
A 53.9� 5.05 5-145 40 B 15.6� 1.58 4-17 16


1988 A 59.6�11.61 15-119 11
B 56.1� 7.19 7-122 20 C 53.5- 4.10 6-142 80










Table 3-11. Origin of new nest owners for the total nests reused during the two breeding seasons studied.



Total Nests Origin of New Resident Year Site Reused From same tree New to tree
1987 Log 28 8 20
A 19 4 15 B 0 0 0
Total 47 12 35


1988 A 8 3 5
B 5 2 3 C 34 14 20
Total 47 19 28


Overall Totals 94 31 63







Table 3-12. Male egg acquisition and hatching success for the 1987 and 1988 breeding seasons.


Year 1987 1988 1987 1988
Site(n) LoQ (63) A (40) B (16) A (12) B (20) C (76) TOTAL(227)

Mean Eggs 21.9 16.5 10.7 22.5 22.9 25.4 21.4 SE 4.2 2.0 3.4 5.3 4.8 1.9 1.3 Range 1-102 0-57.5 0-50.6 0-50.8 0-96.7 0-69.6 0-102

Eggs/d 0.3 0.5 0.7 0.5 0.4 0.7 0.5 SE 0.05 0.12 0.21 0.15 0.12 0.12 0.06 Range 0-1.3 0-3.4 0-2.7 0-1.8 0-2.4 0-8.3 0-8.3

Mean Juv. 2.9 4.1 10.1 47.4 27.5 27.7 18.5 SE 1.2 0.9 4.1 20.9 6.7 4.2 2.3
Range 0-34 0-15 0-50 0-211 0-95 0-152 0-211

Juv./d 0.04 0.08 0.57 0.52 0.42 0.39 0.30 SE 0.01 0.02 0.21 0.17 0.12 0.05 0.03 Range 0-2.7 0-0.5 0-2.4 0-1.8 0-1.9 0-2 0-2.4

Success of males was computed first as the mean number of eggs guarded or mean number of juveniles produced over the residency time of males at nests (=mean eggs and mean juv.); second, as the total number of eggs or total number of juveniles over total number of days as a nest owner(= eggs/d and juv/d)





73







Table 3-13. Correlation between the number of eggs males acquired and the number of juveniles produced for the two breeding seasons studied (Pearson correlation coefficients).


Year Site r D n(males)

1987 Log 0.67 <0.001 63
A 0.40 0.01 40 B 0.87 0.001 16

1988 A 0.77 0.006 12
B 0.23 0.24 20 C 0.58 0.0001 76

Overall Totals 0.47 <0.001 227





74







Table 3-14. Behavior patterns shown (total number of
occurrences, except for time out of nest) by successful and unsuccessful males during 48 h of accumulated focal animal observations at the Log Site. Each male was observed for a
total of 6 h in two, 3-h blocks.




Type of Male
Activity Successful Unsuccessful (n=4) (n=4)
Out of nest 72.3 � 30.4 47.0 � 18.7 (x � SE min)

Brings materials 1 1 to nests

Cleans nest 3 1 Chases other
individuals 0 1 Repairs walls 3 0 Copulations 1 0





75









"1 O Expected Ovipositions 12 -I Observed Ovipositions

m 10

8



O 4 "





No. of Ovipositional 30 d




Figure 3-1. Frequency of ovipositions (No. of new eggs/3) observed over 30 d in 50 nests and the expected Poisson frequencies.






76













O Nests with less than 3 owners (n=14) 50- E Nests with 3 or more owners (n=14)

40

2 30

20

10 0*
Eggs Juveniles





Figure 3-2. Number of eggs and juveniles in nests that had three or more resident males through the 1988 season compared
to nests that had one or two owners. Sample nests were derived from the C Site.






77














G) 200
a
4J 160






0 m


0 20 40 60 80 100
No. Eggs Accumulated




Figure 3-3. Number of juveniles produced at a nest as
a function of the number of eggs accumulated in the nest throughout the male's residency, for both seasons studied
(r=0.47, n=201, p<0.01).






78












M Early season (n=7) M 1 Mid season (n=10) o 1.0 -~ Late season (n=7)

4 0.8


S0.6


0.4


0.2
0 0 5 10 15 20 25 30 Days



Figure 3-4. The probability of a male getting his first egg. Males were followed over 30 d during early, mid and late season of 1988.





79









e 4 [] Before Switching SE3 After Switching

a 3 8
7- 7
-7




SI10



SM G/G G/B B/G B/B 4 Male and Nest Site


Figure 3-5. Changes in male success, measured as the mean number of resident females associated with males at good and bad nest sites after the switching of previously successful and unsuccessful males. Success of males was monitored 15 d
before and 15 d after the switching. Sample sizes are indicated above the columns. G/G=males from good sites
switched to good sites, G/B=males from good sites switched to bad sites, B/G=males from bad sites switched to good sites,
B/B=males from bad sites switched to bad sites.












CHAPTER 4
FEMALE ASSOCIATIONS WITH NESTS AND FEMALE SUCCESS




The reproductive success of both females and males is influenced by where females oviposit. How many and which nests to visit, whether to eat eggs or not, whether or not to court and copulate with a particular male or to leave a nest, and, if they stay in a nest site, for how long, are some of the decisions involved in determining which nests females use. There are different trade-offs associated with each of these decisions (Dunbar 1983). The decision rules may change according to the interaction of females with males, the physiological condition of the female, and ecological variables related to the nests and the nest sites.

Understanding female mating decisions is crucial for

understanding the origin of the mating system (Wittenberger 1983). The patterns of female decisions are based on the relative value of material and genetic benefits and the correlation in the quality of benefits a male may offer (Borgia 1979). The importance of female versus male mate choice and what are the criteria for choice are interesting for species with male parental care (Thornhill 1979). They may provide insight into the operation of sexual selection in a particular system. This chapter examines the association of females with males/nests and explores the relationships




80





81



between female behavior and the reproductive success of females and males.

The study of female behavior was difficult. When trying to understand the behavior of females I found it was easier to follow and quantify the behavior of females who remained at nest sites than the behavior of those that did not. For isolated nests, the assumption that females remaining at the nest site were laying eggs at that nest appeared reasonable. I assigned eggs accumulated in the nest to the resident female(s). For nest sites with more than one nest per tree, I was not completely sure which female was ovipositing where, unless I directly observed the copulation and oviposition events. Consequently, I have limited most data analyses of female behavior to nest sites with one nest per tree. Some analyses are presented however from the Log study site since it permits me to compare the behavior of females under the same nest and male availability and general site conditions.


Types of Visits and Female Condition


Females of Z. albomarainis mature a few eggs at a time

throughout the breeding season. When ready to lay eggs, they enter nests, initiate courtship, oviposit and leave the male guarding the eggs (Mora 1987). They wander from tree to tree visiting nests or they remain associated with nest sites for variable lengths of time.





82



Females may visit nests for different reasons: to mate, to assess the male and the nest, or to feed on eggs. I have categorized and described in a previous study female nest visitation in the following way (Mora 1987): (a) in-and-out visits (ca. 70%, n=137 visits observed in 1985), (b) courtship (successful or not) and copulation visits (28%, n=137), and (c) visits leading to male chasing the female off the nest (2%, n=137). One reason for these differences in female behavior and male chasing of some females may be differences in the reproductive condition of females.

If correct, I would expect that females chased from nests would have fewer mature eggs than those that successfully courted a male. These rejected females entered nests only to feed on eggs. Or it may have been that males could assess female fecundity and rejected courtship from females not likely to oviposit in his nest right after copulation. This explanation, however, does not have clear predictions either for the function or for the reproductive condition of females performing in-and-out visits. For example, females assessing the males and nests might or might not have been ready to lay eggs. Did males chase off and bite females of lower than average fecundity? What was the reproductive status of females visiting nests and did it differ for females performing different kinds of visits?





83




Methods


During the breeding season of 1987, I accumulated with the help of my field assistant, more than 120 h of focalfemale observations. We performed the observations opportunistically, selecting nest sites with good visibility and with guarding males at the nests. We collected females that were carrying out different kinds of behaviors. We dissected those females in the laboratory for the number of mature eggs, that is, eggs that were the size and color of those being oviposited.

Females were divided into five behavioral categories:

(a) Females remaining inside the nest for more than one hour (n=13). Females entered the nest and stayed motionless for prolonged periods of time, which were always more than one hour and extended for up to three hours, the maximum time allocated to the observation of a single female. I considered this behavior part of the courtship sequence in a previous study (Mora 1990). (b) Females leaving the nest after ovipositing (n=8). These females were observed ovipositing, and were collected as they were leaving the nest. (c) Females going in-and-out of nests (n=7). These females entered a nest, tapped the substrate and left. (d) Females chased off the nests and/or bitten by the male (n=6).

These females were collected as the male was running after them. Two of these females were bitten after laying one egg.





84



(e) Wandering females (n=12). These were females collected from trees that had no visible nests. Results


Three kinds of eggs were found in the dissected females. Large numbers of small white eggs were found close to ovaries; their numbers were too large to count accurately, but varied between 6 and >200. A small number of medium sized yellowish eggs were found along the oviduct; these were not found in all females and their numbers never exceeded six. Large yellow eggs were found at the distal portion of the oviduct and filled most of the abdominal cavity. We considered these mature eggs ready for fertilization since they were the size and color observed during oviposition. The following analysis refers to the latter category of eggs.

Females from different behavioral categories had

different numbers of mature eggs (Kruskal-Wallis one-way Anova, H=15.75, d.f.=4, p<0.05, Figure 4-1). There was a gradation of the mean number of eggs ready for oviposition within the females dissected. At one extreme, the females remaining inside a nest for extended periods had the most eggs available for fertilization. On the other end, wandering females had the fewest. Females leaving the nests after laying eggs in them had the second highest number of eggs ready for oviposition, followed by females bitten or chased off the nests. Females performing in-and-out visits





85



had the least number of mature eggs of all females visiting nests (Figure 4-1).


Discussion


These data showed that the physiological condition of females influenceS the kinds of visits they perform. They support the assumption that females arriving at nest sites, entering nests, and staying at nest sites were all females which potentially could copulate and oviposit. Those wandering on trees with no nests were not contributing eggs to the population at that moment. I predicted that chased off females were those entering nests only to eat eggs and therefore would have fewer eggs than those not chased. This was not true. Females that finished laying eggs still had a high number of mature eggs and this may explain why females remained associated with nests for prolonged periods. This observation also supports the assumption that those females that remained at the nest sites for more than three days probably oviposited at least once in that nest. Females sitting at nests were probably waiting to copulate and oviposit whereas wandering females were or may have been ripening eggs rather than looking for oviposition sites. The latter point raises the interesting question of why females may choose to ripen eggs away from nest sites.

It is difficult to interpret the relationship between female nest visitation behavior and reproductive condition.





86



One cannot predict the kind of visit a female may perform to a nest from her reproductive condition, except for the females with very few or very many eggs (at the extremes of the distribution).

One male behavior related to the female reproductive condition is intriguing. Why did males chase off females with ripe eggs? The possible explanations for such behavior are that (a) males chased off females that visited the nests to feed on eggs, (b) males rejected females because additional eggs may be detrimental to those already present in the nest and (c) males rejected females on bases other than their fecundity at the time of the visit. I do not have information on the previous association of the rejected females to the nests from which they were chased off, but I would expect that females that have not oviposited in the nest would be more likely to feed on the eggs present than resident females would. Male mate rejection of unfamiliar females may explain the association patterns of females with nests discussed in the following sections. Since the hatching success of nests was correlated with the number of eggs a male gets to guard (Figure 3-3, Table 3-13), there does not seem to be any reason for guarding males to avoid getting additional eggs. Hence, the detrimental effects of receiving additional eggs may not be the reason for males to reject females. In species where paternal care occurs, male mate choice is expected to become more important than for species in which males provide low parental investment





87



(Ridley 1978; Searcy 1982). Male mate choice could be based on a number of traits not measured in this study that may indicate the genetic quality of the females (Petrie 1983). Whether or not Z. albomarainis males are choosing those females with "better genes" is a question that will require laboratory mate choice experiments.


Patterns of Female Associations



Operational Sex Ratios


I estimated the sex ratio in the nesting population to be 1 male:3 females(393/1282). There was a greater proportion of females in relation to males in the 1987 than in the 1988 nesting season (Table 4-1). The proportion of males to females varied between study sites but these differences were not consistent between years. The sex ratio of nesting adults was most biased in the A Site, with a proportion of six females for every male.

These differences in sex ratios between sites suggest a patchy distribution of individuals. Consequently, not all males face the same level of competition to attract females and the competition for nests among females may vary from one patch in the forest to another.

Female longevity. Females can survive to a second breeding season; three females from 1987 were observed visiting nests in 1988.





88




Freauencv and Patterns of Visits


Most females observed around nests had eggs ready for

fertilization and oviposition and females had the capacity to travel from tree to tree in search of males and nests. Since nests with guarding males were abundant in the study area, one may assume that females arriving at nest sites were looking for oviposition sites and therefore were faced with a series of decisions: (a) how many nests to visit, (b) which nests to visit, (c) stay or leave a nest site, and (d) if they stay, for how long. Following, I will describe the overall population patterns of nest visitation.

Methods. The best way to quantify the number of nests

visited by females would have been to follow the movements of marked individuals through the entire study area for extended periods of time. I have found this practically impossible at the forest sites. Once a female left a marked tree, the chances of seeing her again were very low. Also, most of the movements were carried out at night when observations were difficult. As such, I chose to estimate the extent of female movements between nests from the rates of female immigration to nest sites considering only trees that had one nest. These movements were recorded as part of the regular population censuses. This estimate is conservative because it is based on only one observation every 2 d per site, and





89



it does not consider movements between trees with multiple nests.

I also analyzed female visits to nests for 48 h of

accumulated observations in 3-hour blocks at the Log Site. Here one may assume that analyzing the female movements from nest to nest may give an accurate estimate of the female tendency to visit several nests, controlling for the associated decision of whether to stay on the tree or not. Following female movements within the same tree has the added advantage that I could observe several females simultaneously under the same conditions of time of day, nest availability and meteorological conditions.

Results. Female movements were remarkable in the

forest: marked individuals were recorded traveling more than 500 m from the nest at which they were marked over a two-day period. The mean rate of female arrival at nest sites was

0.52 female/day (SE=0.27, range=0-3.5, n=145).

The mean total number of different females that visited an individual nest site was 5.11 (SE=0.77, range=0-23 females, n=145) for all forest sites. Twenty-two percent of all nest sites never received any female visits and less than four percent received visits from more than 20 different females during the breeding season (Figure 4-2).

At the Log Site, individual females visited a mean of 0.87 nest/day (1.73 nests over 48 h in 3-h blocks, SE=0.18, n=77); 25% visited none, and one female visited six nests. The frequency of visits to nests at the Log Site was not





90



significantly different from that at the forest sites (X2=0.18, d.f.=l, p>0.05).

Discussion. Under the same conditions of nest

availability and in the same nesting area (as in the Log Site), females showed individual differences in mobility between nests. However, the extent and consistency of these differences was not examined. This pattern of nest visitation may have been the result of different rates of movement between females or by the different ability of males/nests to attract visits.

Individual differences in mobility may also be present in females visiting nests at the forest sites. Forest females, however, must travel greater distances than females at sites with multiple nests. The cost of visiting nests may vary for females on trees with multiple nests as compared to females on trees with a single nest. It would be desirable to measure experimentally the differential survivorship for females in a log and a forest site or to measure the time between ovipositions for females that visit several nests in the same tree as compared with females who visited several trees. The difficulty of following females from tree to tree at the forest sites made such analyses practically impossible.

There were nests for which I had no recorded female visits for the entire breeding season and nests that attracted a large number of females. Is this a random





91



pattern or are females attracted differentially to certain nests?


Do Females Visit at Random?


Some nests in the forest site received no female visits during the nesting season while others received up to 23 visits (Figure 4-2). This variation is important because it suggests the existence of female preferences. Such a pattern of female visits may have been generated because females are indeed attracted to some nests more than others, or it might have been generated randomly. Are there in fact nests receiving too many or too few visits as compared to the visitation frequency expected by chance alone?

Methods. I randomly selected from the forest sites 50

nests occurring singly at trees. I extracted from the census data the number of different females that visited the nests for a 30 d period in the 1988 season. I considered a visit the appearance of an unmarked female or of a marked female new to the tree. I calculated the distribution of the number of visits per nest and compared that to an expected Poisson distribution. The two distributions were compared with a two-tailed, Chi-squared test with an alpha level of 0.05.

Results. The mean number of females that visited a nest over a 30 d period for 50 nests was 2.6 females (SE=1.26, range 0-9, n=129, Figure 4-3). The distribution of female





92



visits was not different from a Poisson distribution (X2=6.07, d.f.=5, p=0.29, Figure 4-3).

Discussion. If females arrived on trees in a non-random pattern, I expected to find too many nests with no visits and more nests than expected with a large number of female visits. The results indicate females visited nests randomly and the proportions of nests with zero or with any number of visits were not different from those expected by chance.

Although this analysis refers only to trees with one

nest, the result was surprising since it shows that females were not attracted to certain nests more than others. Female movements did not follow directional or biased patterns that would have produced clumping of the visits. A random pattern of nest visitation may be generated by random encounters of females with trees.

This may partially explain why males remained at nest sites even when they had received few or no female visits (Chapter 3). Since arrival of females at trees is random, a better estimate of female choice may be derived from the female residency time at nest sites and from their rates of oviposition than from their rates of visits to nest sites. Length of Female Associations to Nests


Female Residency. I assume that females arriving at a nest site could remain at a given nest site or leave to search for another. The possible association of females to




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ETQWY6WEY_9INDJO INGEST_TIME 2015-04-13T19:09:52Z PACKAGE AA00029938_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

SITE-BASED MATING SYSTEM IN A TROPICAL HARVESTMAN BY GISELLE MORA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1991

PAGE 2

ACKNOWLEDGEMENTS I have benefited during the writing of this thesis and my graduate studies from the encouragement, comments and criticisms of many people. It has been with their help that I was able to successfully complete this project. I would like to thank the members of my committee, H. Jane Brockmann, Jonathan Reiskind, John Sivinski, John Anderson and William Eberhard for their valuable comments throughout the various stages of my research and for their criticisms during the final revision of this work. Special thanks go to my adviser H. Jane Brockmann. She has contributed immensely to organizing the material, to improving the quality of this thesis and in many other aspects of my education. Linda Fink has also contributed with ideas, comments and has been a supportive friend throughout. Julio Arias helped with statistical analyses. The Department of Zoology, University of Florida, has been a stimulating and supportive place for conducting my graduate studies. I would like to thank Dr. Frank Nordlie, Mrs. Lynda Everitt and Mrs. Carol Binello for logistic support . I have also received much support and encouragement while at the Smithsonian Tropical Research Institute in

PAGE 3

Panamd. Drs. Egbert Leigh and Donald Windsor acted as Smithsonian advisors of this project. Egbert Leigh has supported this research since 1983 and gave me the confidence for continuing during the darkest moments of my field work. I thank him for not letting me forget the value of natural history. Hector Barrios assisted me in the field during the 1987 season. Argelis Rom^n and Georgina de Alba provided logistic support and help with many matters during my stay in Panama. Rayneldo Urrutia coordinated many administrative aspects of my residence at BCI. I overlapped with many researchers during my residency at BCI; many of them contributed with observations, criticisms and friendship. In special, I appreciate the encouragement I always received from Phil DeVries. The support of my family and friends have brightened my years as a student. I thank my entire family for their support and love. In special, I thank my parents for always letting me try to be all what I thought I could be. I also thank Juan Carlos Vargas, Julio Arias, Clara Sotelo, and Catherine Langtimm for their friendship and support especially during the writing of this thesis. My husband Henry put up with me, fed me, and kept his and my spirits high when the writing of this thesis got rough. I thank him for never losing his sense of humor and for helping me finish. I have been supported by a Noyes Predoctoral Fellowship from the Smithsonian Tropical Research Institute, by teaching

PAGE 4

and research assistantships from the Department of Zoology, University of Florida and by a scholarship from the Consejo Nacional para Investigaciones Cientificas y Tecnol6gicas, Costa Rica. iv

PAGE 5

TABLE OF CONTENTS ACKNOWLEDGEMENTS ii ABSTRACT vii INTRODUCTION 1 BACKGROUND AND METHODS 5 Natural History and Reproductive Biology of Opiliones 5 Reproductive Biology of Zyaopachylus albomarainis 12 Study Site 14 Study Population 16 General Methodology 16 Definitions 18 NESTS AND MALE SUCCESS 22 Nests and Male Success: General Descriptions 23 Patterns of Nest Use 28 Patterns of Nest Success 32 Male Life History and Behavior: General Descriptions 39 Components of Male Success 43 Why do Males Abandon Nests? 50 Nests or Males? A Male Switching Experiment 52 Discussion: Patterns of Nest and Male Success 55 FEMALE ASSOCIATIONS WITH NESTS AND FEMALE SUCCESS 80 Types of Visits and Female Condition 81 Patterns of Female Associations 87 Female Associations and Nest/Male Measurements 97 Components of Female Success 99 Discussion: Females Association with Nests 103 DO FEMALES ASSOCIATE WITH MALES OR NESTS? 120 Methods 121 Results: Female Residency Before and After the Switchings 123 Discussion: Females Associations With Nests 123 v

PAGE 6

THE NEST SITES 127 Rainfall 129 Fungus 138 Predators 143 What Accounts for the Differences in Nest Success? 146 SITE-BASED MATING SYSTEM OF Zyaopachvlus albomarainis 162 The Mating System of Z.albomarainis 165 The Evolution of the Site-based Mating System of Z.albomarainis 176 LITERATURE CITED 181 BIOGRAPHICAL SKETCH 189 vi

PAGE 7

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SITE-BASED MATING SYSTEM IN A TROPICAL HARVESTMAN By Giselle Mora December, 1991 Chairperson: H. Jane Brockmann Cochair: Jonathan Reiskind Major Department: Zoology Zyaopachylus albomarainis Chamberlin (Arachnida, Opiliones: Gonyleptidae) , a tropical harvestman, is the only arachnid exhibiting paternal care. Males construct mud nests on trees that females visit to court the resident male and oviposit. I conducted a two-year field study on Barro Colorado Island, Panama to identify and evaluate components of reproductive success and to characterize the mating system of a natural population of this species. The mating system of these harvestmen is polygynandrous, with females laying eggs in several male nests and males collecting eggs from several females. Females mature a few eggs at a time throughout a breeding season lasting from June to December. Females exhibited a random pattern of nest visitation but the vii

PAGE 8

distribution of oviposit ions was not random. This mating system resembles those of many substrate-nesting, eggguarding fish. It is site-based since both male and female reproductive success depends upon their association with successful nesting sites. Breeding success of males was measured by (a) the number of eggs acquired during his residency (b) the number of juveniles hatched, and (c) the number of resident females associated with the male and his nest. The ability of males to acquire and hold a nest at a good nesting site accounted for much of the variability in male breeding success. Male reproductive success was not correlated with body size. Switching experiments revealed that resident females continued to associate with the site regardless of the identity of the male at that site. Ecological variables, including rain, occurrence of fungal attack and incidence of predation may account for the success of some nesting sites over others. I hypothesize that this mating system evolved from no parental care. A limited availability of safe oviposition sites may account for the evolution of male territoriality, male egg-guarding and female attraction and fidelity to nesting sites. viii

PAGE 9

CHAPTER 1 INTRODUCTION The harvestman Zyaopachvlus albomarainis (Opiliones: Gonyleptidae) is the only arachnid exhibiting paternal care (Mora 1990) . Their reproductive behavior is remarkable for an arachnid. Males guard eggs in nests made out of mud and tree bark. Besides building a nest, males take over or use abandoned nests. Females visit nests, court males, oviposit, and leave the eggs for the male to guard. These behaviors are unique among opilionids. I present here the results of a field study that focuses on the variation in reproductive success of males and females and the factors that contribute to such variation. The objectives of the study were to identify and evaluate components of reproductive success and to characterize the mating system of a population from Barro Colorado Island, Panama. My ultimate goal was to gain understanding of selective pressures that have shaped this mating system. The importance of this study lies in documenting the reproductive behavior of the only arachnid with paternal care. It also allows a test of current ideas on the influence of increased paternal investment on male and female mate choice and relative variation in reproductive success. 1

PAGE 10

2 The first model explaining the factors contributing to the variation in reproductive success was proposed by Darwin (1871) . He recognized two types of selective forces acting on the context of reproduction. Intrasexual selection is the competition within one sex for individuals of the opposite sex and intersexual selection is the preferential choice for mating partners by one sex. Darwin considered intrasexual selection to be more common and more pronounced in males and intersexual selection to be an attribute of females. Although this distinction has proven useful, most biologists now recognize the difficulty of separating the two (Maynard Smith 1987) . Trivers (1972) elaborated Darwin's idea by arguing that the sex with the greatest parental investment will become a limited resource for the opposite sex. He recognized as parental investment those activities of an animal that increase the probabilities of present offspring survival and reproduction at the expense of the parent's ability to have additional offspring. Consequently, the sex investing the least will compete among themselves in order to mate with individuals of the sex investing the most. Since females generally produce fewer and energetically more expensive gametes than males, their parental investment is often greater. This produces the general pattern of male-male competition and female mate choice. When a male invests, he may become a limited resource for females and the pattern may be reversed. Also, phylogenetic and ecological factors will

PAGE 11

3 produce variation in the degree of female competition for males and male choosiness. Such factors have been investigated for some species (Petrie 1983), but they have not been evaluated in any harvestman. Evaluating the factors that contribute to the variation in reproductive success of males and females is necessary to understand any mating system. For Z. albomarainis males, among the factors that may contribute to individual variation in reproductive success are age, size, and the ability to acquire and retain a nest. Since egg hatching success depends on the presence of a male (Mora 1990) , differences in paternal behavior may account for variance in reproductive success. Paternal behavior may also be a criterion by which females assess male quality (Petrie 1983). Female reproductive success is associated with the success of the nest in which she oviposits. Little was known about the behavioral ecology of female reproduction prior to this study. Physiological condition and attachment to nesting sites will be explored as contributors to the variation in female reproductive success. These attributes and behaviors of males and females have been recognized as important contributors to reproductive success in many species (Clutton-Brock 1988) but have never been evaluated for this harvestman . Since males without nests do not acquire mates, I evaluate in Chapter 3 the patterns of nest use and nest success, and present experimental evidence showing the

PAGE 12

4 contribution of nests to the male's reproductive success. For females, an important reproductive choice is whether to associate with a male and his nest or to search for additional mates. I discuss the nature and the patterns of these associations in Chapter 4. Whether females associate with males or with nests was tested with a male switching experiment; I present these results in Chapter 5. In addition, I discuss the ecological setting of the nesting sites in Chapter 6, and characterize the mating system and the components of male and female reproductive success in Chapter 7. I argue that this is a nest site-based mating system; that is, a system in which both male and female behavior and reproductive success depend upon their association with good nesting sites. Finally, I discuss a possible scenario for the evolution of this mating system in Chapter 7 .

PAGE 13

CHAPTER 2 BACKGROUND AND METHODS Natural History and Reproductive Biolocfv of Opiliones Reproduction in Zyaopachylus albomarginis Chamberlin includes a number of behaviors unusual or unique among Arthropoda. However, this apparent exceptionality may be simply due to our lack of knowledge of natural history of many arthropod groups. Despite the fact that the arthropods constitute approximately four-fifths of the world's animals (Savory 1977), most are poorly known. Our ignorance is particularly evident in the Arachnida, a large and diverse class composed of 12 living and five extinct orders. Of these, the Araneae (spiders) and Acari (mites and ticks) are the best studied. Field and laboratory studies on the ecology and behavior of the rest of the arachnid groups are practically nonexistent and limited mostly to temperate species. Most studies on opiliones have been on species of the suborder Palpatores perhaps because these are common in North America and Europe where a longer tradition of ecological studies exists (for a review see Savory 1977; Hillyard and Sankey 1989). After Acari and Araneae, the Opiliones are the third largest order within the Arachnida. They are distinguished 5

PAGE 14

6 from other arthropods by having a body divided into two parts (an anterior cephalothorax, prosoma, and a posterior abdomen, opisthosoma) ; a pair of chelicerae, a pair of pedipalps and four pairs of legs, no antennae or mandibles, and simple eyes. The order is also characterized by the presence of a pair of odoriferous or repugnatorial glands that are usually situated at the bases of the first or second pairs of legs. The cuticle of the body and appendages in opiliones present a range of hairs, setae, spines, denticles, and tubercles that form the "armature". These projections can sometimes become very elaborate (Rambla 1975; Hillyard and Sankey 1989). Opiliones, unlike other arachnids, respire by means of tracheae . The group is estimated to contain between 3 500 and 5000 species (Rambla 1975; Hillyard and Sankey 1989) . Species of harvestmen generally do not have an extensive distribution and have little capacity for long range dispersal (Savory 1977). Three suborders are presently recognized (Hillyard and Sankey 1989): the Cyphophthalmi is a very small and primitive group with approximately 50 known species; members of this group resemble mites. The Laniatores or short-legged harvestman are characterized by having the coxae of the first three pairs of legs meeting at the midline, and the pedipalps having a characteristic large terminal claw. It is the dominant group of harvestmen in the tropics and subtropics. Cosmetidae and Gonyleptidae are the largest families and are most diverse in South and Central America where they

PAGE 15

7 constitute one-fourth of the suborder; albomarainis is a laniatorid gonyleptid. The suborder Palpatores are the very familiar daddy-long-legs that are common in the north temperate zone; they are recognized because the coxae of the first three pairs of legs do not meet at the mid-line. Phylogenetically , Opiliones are most closely related to Acari; notostigmatic mites are believed to be very closely related to cyphophthalmid opilionids (Savory 1977). Natural History Like many other arachnids, opilionids are long lived. Juberthie (1967) recorded a cycle of nine years for a cyphophthalmid. However, most of the temperate species have an annual life cycle (Edgar and Yuan 1968) overwintering as near-adults or as eggs (Hillyard and Sankey 1989) . Most tropical species probably have life cycles of more than a year with subsequent overlapping of generations (personal observation) . It may take up to eight molts to attain adulthood . The ecology of the early stages of opilionids is a mystery for most species. In some Leiobunum species, young individuals ascend to the tree and shrub canopy as they mature (Edgar 1971) . Tropical species may spend juvenile phases in the forest litter or in the canopy as revealed by evaluation of museum collections (personal observation) , Adult individuals commonly form large aggregations, familiar

PAGE 16

8 to all, but little studied and understood (Coddington ^ al . . 1990) . There have also been observations of territoriality with male-male antagonistic behavior reported for some species of Leiobunum (Edgar 1971) . Because of their secretive lives and inconspicuous appearance, they have few known important predators. Nevertheless, predation by frogs, lizards, birds, mammals, centipedes, spiders and several insects has been reported (CloudsleyThompson 1968) . Most species are omnivores, and opilionids have been recorded feeding on other harvestmen, snails, worms, millipedes, woodlice, earwigs, flies, bodies of dead ants and beetles. In addition, they nibble on gills of fungi, dead moles and mice, and bruised and fallen fruit (CloudsleyThompson 1968; Savory 1977) . The diet of Opiliones differs from that of spiders. Although the latter are limited to predigested prey juices, the former chew on soft food. Cannibalism, especially among juveniles, is common (Edgar 1971) . Changes in temperature and humidity have been suggested as the major factors of mortality for the group. They are more susceptible to dehydration than most other arachnids. In the laboratory, harvestmen have great difficulty molting when the temperature and humidity are not adequate (Hillyard and Sankey 1989 ) . Very little is known about the behavior of natural populations of opilionids. Most opilionids are nocturnal.

PAGE 17

9 The standard behaviors reported for harvestmen are negative phototaxis (movement away from light) , thigmotaxis (tendency to press the body against the substrate) , hydrotaxis (sensitivity to presence of water) , and rheotaxis (response to movements) . They can autotomize legs when captured, exert secretions from repugnatorial glands, and show catalepsy (faking death) when disturbed (Hillyard and Sankey 1989). Reproductive Biology Opilionid reproductive biology, like that of the Acari, is unusual among Arachnida in their direct use of an intromittent organ during copulation (Berland 1949; Thomas and Zeh 1984) . Copulation, according to most observers, "appears to be a casual affair occurring freely and frequently" (Hillyard and Sankey 1989). Studies of tropical laniatores may prove this statement wrong. Reproduction in Cyphophthalmi involves the production of a spermatophore; therefore, copulation in this group does not occur as in other opilionids (Hillyard and Sankey 1989) . Generally there is no marked sexual dimorphism in opilionids; however, in some species there are notable differences in palpi and abdominal sizes (Thomas and Zeh 1984; Hillyard and Sankey 1989). The reproductive system of the females is composed of a horseshoe-shaped ovary, with its anterior ends joined to a median oviduct that leads to the ovipositor which is a broad and mobile tube made up of

PAGE 18

10 membranous rings. The reproductive system of males is a Utube shaped testis, with a median vas deferens leading to the penis which is more rigid than the ovipositor. Penis morphology is a species-specific character (Berland 1949; Savory 1977) . Both ovipositor and penis lie behind the genital operculum, a fold of tissue that covers the reproductive organs when at rest. The genital opening, although it is derived from abdominal segments, is located between the leg coxae. During copulation, partners face each other while the male grasps the female with his pedipalps and extends the second pair of legs outwards. The penis is then extruded and the female may use her pedipalps and chelicerae to guide it to the genital opening (Juberthie 1965) . The duration of copulation can vary from a few seconds to several minutes depending on the species and the behavior of the female. Repeated copulation with a female seems to be the rule for the palpatorid species for which information exists (Edgar 1971; Hillyard and Sankey 1989) . The only existing detailed description of courtship and copulation for a laniatorid comes from albomarainis (Mora 1987). That study reports a much more complicated courtship and copulatory behavior than previously thought. Generally, the female approaches the nest, enters and taps the substrate. She then alternates, in variable order, between tapping the partner, moving around the nest, facing the partner and remaining quiscent until the male starts tapping the female. He then moves to her side.

PAGE 19

11 her back, and underneath her as he taps the substrate. This continues until the partners face each other and the female extends her chelicerae and palps, grabbing the male by the cephalothorax and pulling him toward her. In this position, the male everts his penis and intromission takes place. Oviposition always follows within 20 min (Mora 1987). Parthenogenesis may be common in opilionids. For some species, males are extremely rare in collections and natural populations. Parthenogenesis has been documented in Phalanaium o pilio and Meaabunus diadema (Phillipson 1959; Edgar 1971) . To further test parthenogenetic capabilities of M. diadema . Phillipson (1959) collected 10 subadult females and kept them in aquaria without males; seven of them produced 13 batches of eggs, all viable. Fertilization occurs during the process of egg laying when the sperm are released from the spermathecae, enabling one sperm cell to enter each egg before the hardening of the chorion and vitelline membrane (Juberthie and Munoz-Cuevas 1971; Juberthie and Manier 1976) . In nature, opilionid eggs are difficult to find and very susceptible to the attack of mold (Edgar 1971; Savory 1977). They are usually spherical and of pale coloration. Generally, they are laid in clusters in places such as damp soil, moist vegetable debris, crevices of bark, under stones or dead wood (Rambla 1975; Hillyard and Sankey 1989). In the 10 British species examined by Phillipson, the number of eggs laid was always only a fraction (21-57%) of the mean maximum number of eggs produced

PAGE 20

12 by the ovaries. This seems to be a common feature of other species (personal observation) . Reproductive Biology of ZYcropachylus albomarainis The reproductive biology of Z. albomarainis is extraordinary among the arachnids. This very small and cryptic species (body approximately 3 . 5 mm long) is known to occur only in Panama. There is no sexual dimorphism in this species. Females are a little larger than males, but the differences are not completely reliable for identification purposes. The species was redescribed from specimens collected on island by Goodnight and Goodnight (1942) . The first report on the nesting behavior of this species was provided by Rodriguez and Guerrero (1976) . Perhaps the most remarkable aspects of this system are the construction of nests by males and the existence of paternal care. No other arachnid is known to have male parental care, although female care is common. Nest construction in the arachnids is found in spiders (e.g. Salticidae, Pisauridae) and in the brooding nests of pseudoscorpions, though it is the female who guards eggs in those cases (Savory 1977). The structure constructed by males of albomarainis . utilizing mud, tree bark, and salivary secretions is unique among arachnids and its evolutionary origin is a mystery. Paternal care in this species takes the form of egg guarding inside circular, open mud nests which are

PAGE 21

13 constructed on standing or falling trees. Males guard the eggs until they hatch, which takes 18-22 days, preventing the growth of fungus and predation by conspecifics and ants. They also repair the nests when damaged by rain, wind, or the passing of larger animals (Mora 1987). Previously I showed that male guarding enhances hatching success. Unguarded eggs are lost because of predation by conspecifics and ants, and because they are washed off by rain, attacked by fungus, or disappear for reasons that have not been established (Mora 1990) . Young Z-x. albomarainis undergo their first two molts in the nests (Mora 1987) . Although males do not care for early instars, nests provide a secluded place for them before they disappear into the forest floor or the tree canopy. I described nest construction behavior in a previous study (Mora 1987, 1990) . Besides building a nest, a male can acquire a nest by taking over an already occupied one or by utilizing abandoned nests. Nests are necessary for males to reproduce since females neither mate nor oviposit outside nests (Mora 1990) . There are, however, nestless males in the population . Females mature a few eggs at a time through the breeding season. When ready to lay eggs, they enter nests, court males, copulate, and then oviposit from 1-3 eggs inside the nest. They leave the eggs for the male to guard (Mora 1987). Females wander on tree trunks, and their behavior was poorly understood prior to this study.

PAGE 22

14 The breeding season on Barro Colorado Island, where I conducted this study, lasts about six months from June to late December. This corresponds to the wet season on the island. The natural history of these animals during the rest of the year is unknown, but they probably live secretive lives on the forest floor, under crevices and on tree trunks until the rains begin. Laniatorid opilionids are known to live several years in captivity (personal observation) and I have observed marked individuals of this species reproduce in two consecutive years. Study Site I studied a natural population of harvestmen on Barro Colorado Island (BCD , Panam^, from June to December of 1987 and 1988. I conducted a previous study of the same population from July to December, 1985 (Mora 1987) . Barro Colorado is a biological reserve protected since 1923 and administered by the Smithsonian Institution since 1946. Its total area is 1500 ha; the island was isolated in 1914 after the Rio Chagres was dammed to form Gatun Lake, which presently makes up the central portion of the Panam^ Canal. The island is a hill rising 137 m above the Gatun Lake. The hilltop is broad and flat, with steep (20-30 degrees) slopes at the edges. About half of the island is covered by young forest, a hundred or more years old, which is still growing back from old agricultural fields. Undisturbed, old

PAGE 23

15 forest (200-400 years) covers most of the rest of the island (Figure 2-1) . Both old and young forest have nearly 60 species of trees per hectare. On average 170 trees over 20 cm diameter at breast height (dbh) can be found per hectare. The height of the canopy in the old forest is 30-40 m (Foster and Brokaw 1982; Leigh 1982). The most conspicuous seasonal changes in the island are associated with rainfall. Therefore, the biology of many organisms is closely tied to rains. The times when the wet and dry season start and stop are perhaps the most important aspects of the biological year on BCI (Rand and Rand 1982) . These are independent of one another. The rainy season usually begins between mid-April to mid-May and lasts until December. Ninety percent of yearly rainfall occurs then (Dietrich , 1982). October and November are the months with the most hours of rain (Rand and Rand 1982). Rain falls usually in short but intense storms. The mean monthly precipitation during the rainy season is 310 mm, as compared to 51 mm during the dry season (data for 50 years, Dietrich , 1982). These wet-season rainstorms cause soils to swell, and in many places to become saturated and discharge water to small ravines scattered throughout the island (Dietrich sX. , 1982). There are, however, dry periods of up to a month during the rainy season, sometimes causing the understory plants to wilt. This short dry season usually occurs during August. Over the past 50 years, the average annual rainfall was 2 600 mm (SD=422 mm, Windsor 1990) .

PAGE 24

The mean monthly temperatures vary little annually. Average annual temperature is 27° C in the open. Relative humidity remains high all year. The monthly relative humidity at midday, as measured by a sling psychometer, normally varies from 75-77% in March or April to 93% by November (Dietrich £L , 1982). Leigh (1982) described the climate, flora and fauna of the island. Study Population The population studied occupies the central plateau of the island. All trees containing nests have been mapped in an area of approximately 500 x 250 m. In 1987 I used three main study sites: (a) Site A consisted of a transect approximately 580 m long and 5 m wide, along short-cut trails at the central plateau of the island, (b) Site B consisted of a linear transect of approximately 400 m along Armour trail in the central plateau and (c) Log Site was a fallen log approximately 12 m long and 1.5m wide, close to mark 15 on Wheeler trail. In 1988 I monitored Sites A and B and added a new plot, Site C, adjacent to Site A, measuring 500 x 154 m (Figure 2-1) . General Methodology Most data I consider in the following chapters were gathered from censuses of nests at the study sites. Trees

PAGE 25

17 that contained nests were individually flagged. For all nests I recorded the height above the ground, the nest diameter (as the average of the maximum and minimum nest width) and the height of walls (the average of maximum and minimum wall height) . I repeated these measurements whenever a new male was found occupying the nest. These measurements were taken following the protocol of Mora (1987). Individuals were color-coded using various combinations of Tester's enamel paint on carapace and legs. Each individual was brought to the laboratory once and measured under a dissecting scope for body length (dorsally from occularium to posterior cephalothoracic spines) , total width (at level of the fourth coxae) , weight (to nearest mg) and total length of tibia + patella of the fourth leg. Every other day all nests in all sites were surveyed, recording the resident male as well as all females and other males present at nest sites, the number of eggs in the nest, number of juveniles, damage to nests, male and female behaviors, and interactions occurring at the time. All censuses were performed between 0800 and 1100 h. Additional data on nest visitation, courtship, and male-male and femalefemale interactions were recorded during about 200 h of opportunistic observations. If a nest was abandoned for more than two weeks, only weekly checks were performed to confirm the status of that nest. After a take-over or a new occupation occurred, the new male was marked and measured. All females present in a

PAGE 26

18 tree with a nest were monitored. Trees with no nests were not censused. I made opportunistic observations of females on trees that did not possess nests. I cannot sex subadult individuals; therefore, I neither marked nor measured them. Definitions Nest owner refers to the resident male of a nest and guards eggs inside it. Nest is the physical structure built by males where eggs are oviposited. Nest site refers to the tree where the nest is located. I expand the term in the last chapter to include an array of physical and biological variables. Resident female is a female who remains at a nest site for more than three consecutive days and assumed to have oviposited in the nest at least once. Residency reflects the outcome of female choice and was used as a measurement of nest and male success. Study site is a plot delimited in the forest where I performed the censuses and most experiments (unless otherwise noted) . Male success is the breeding success of individuals, as measured by the number of eggs a male acquires, the number of eggs hatched at his nest and the number of resident females associated with the male during his nest tenure.

PAGE 27

19 Nest success refers to the number of eggs acquired and hatched, and the number of resident females associated with the nest for the total time a nest is occupied during the breeding season. Wandering female is a female not associated with a nest site. These were females found on trees with no nests. It also refers to females visiting nest sites and remaining there only 1-2 days.

PAGE 28

0) u o 4-1 4-) T5 0) 4J to -H O O CO to CO iJ 03 (1) a 4-1 o >1 i-> -H CO C! (U -O 0) o c o •H 4-) (0 s •H 4-> CO Q) I •H 0) O jJ -H a o e (0 0) ^^ to :3 to JJ to rH (J) to O a TS 4J rH 4.) Xi 13 J-> O to S-i u G 0) u c •H o (0 (D to O CO u to ^ 3: to PQ CO CO 0) CT\ a o a o •H J-> rH •H o a s O 0 U 13 0) JJ a •H 0) !h Q) ^ CO i JJ . O 00 H CTi JJ rH to — JJ Dl 03 0) X > O ^^ m to o cn Q) )H (0 u to a; -H S-i (U JJ M (0 a to JJ PQ 0) jJ .< to u CO -H 0) JJ c -H -H CO

PAGE 29

21

PAGE 30

CHAPTER 3 NESTS AND MALE SUCCESS The most remarkable aspect of the reproductive biology of the tropical harvestman, Zyaopachylus albomarainis is that males construct nests in which they guard eggs. Males build and repair nests with mud and tree bark, defend them, usurp them, and use abandoned nests (Mora 1987). The mating behavior of both males and females is strongly associated with nests: mature males and females spend the entire breeding season associated with them. I present here descriptions of the patterns of nest use and nest success and discuss how those patterns relate to the reproductive success of males. This identification of what contributes to the variation in male reproductive success is a first step toward understanding the evolution of male egg-guarding and the mating system of this species. In this chapter, I will evaluate a male's reproductive success based on measurements of three components of his breeding success: male nest occupation, which measures the male's capacity to acquire and retain a nest; male egg acquisition, which measures a male's ability to attract mates, and egg hatching success, which measures the success of his paternal behavior. The number of females that associate with the males and their nests, and the length and 22

PAGE 31

23 nature of these associations are also important components of the male's and the nest's success. These aspects are discussed in Chapter 4. The difficulty of distinguishing male attributes from nest qualities when evaluating male success was a problem encountered throughout this study. The problem arises because females will not mate or oviposit outside nests. Hence, the success of a male is determined by his ability to acquire a nest, and the quality of the nest and nest site. At the end of this chapter I present results from an experiment that attempted to partition out the relative contribution to reproductive success of males and their nests . Nests and Male Success: General Descriptions Nest Distribution and the Sampled Population Nests and individuals of albomarainis were abundant on Barro Colorado Island. I marked 223 nests and monitored a total of 158 nests. I followed 1031 individuals of the 1675 marked during the 1987 and 1988 breeding seasons (Tables 3-1; 3-2) . The distribution of nests on the island was not continuous. Rather, nests were concentrated in the upper plateau where there was a combination of old and tall, young forest. Most areas with young forest, scrubby secondary

PAGE 32

24 growth and those adjacent to the island edges had very few or no nests (Figure 2-1) . There were additional areas on the island, of similar forest composition to the one of highest nest densities, that were not utilized as nesting areas. I consider all sites except the Log Site (a fallen log along a trail) , as forest sites, with nests both close to and away from trails. Nests were not more abundant on any particular tree species. They were rarely found on small trees; only 3 4 out of 223 nests monitored were built on trees of less than 40 cm dbh (x^=107.7, p<0.001, df=l) . These observations suggest that males established nests on mature trees, which, besides having a greater girth, perhaps provided greater microhabitat complexity by having more epiphytes, more food sources, and more complex architecture than young trees. Contrary to previous observations (Mora 1987), pooled data from 1987 and 1988 showed no difference in the number of nests found on buttressed and unbuttressed trees (X^=0.74, df=l, n=223, p>0.05). However, there were more aggregations of three or more nests on buttressed trees than on smooth ones (Table 3-3). This difference may be an artifact due to the fact that buttressed trees are generally larger in surface area than unbuttressed ones. Based on my observations on the main causes of nest destruction and egg mortality (Mora 1990, this study Chapter 6), some variables that may account for the differential use of nest sites are the amount of rainfall running down the

PAGE 33

25 tree trunks, the probability of fungal infestations, and predator attacks at the nest site. Differences in canopycover may produce differences in the amount of water running down the trunks. This flow of water not only destroyed eggs when it accumulated in the nest, but it also washed off the nest walls and nest floor. Nest sites may also differ in their probabilities of experiencing attack by fungus or egg predators. Males clean the nests and are able to deter the damage of fungus, but probably at a high cost (see Chapter 6) . The main egg predators are conspecifics (females and males) , several species of ants, and an unidentified species of flatworm. Several characteristics of the nest (height, orientation) may increase its likelihood of an attack and several characteristics of the sites (humidity, temperature, shade, forest composition) may attract some kinds of predators. The patterns of nest and egg damage due to these variables are discussed in Chapter 6. Measure of Success The breeding success of an organism has various components. Among them, survival to breeding age, reproductive life span, fecundity, mating success and offspring survival have been proposed as important sources of variability in success among organisms of a given population (Clutton-Brock 1988). I cannot age individuals and have not

PAGE 34

26 made observations between nesting seasons; my measures of success are derived from the observations and experiments on individuals present during the nesting season. This will lead to an overestimate of the average success of individuals in the population since it does not take into account the proportion of individuals that never reach reproductive age or fail to join the nesting population (Fincke 1988) . I measured components of male success related to their preand post-mating success. The pre-mating success of a male includes survival to breeding age, success in intrasexual competition, and mating success. Success in intrasexual competition refers to building, taking-over and finding abandoned nests, and at keeping a nest. Mating success is derived from the male's chances of receiving female visits and securing copulations at the nest. I estimate premating success from the male's ability to acquire and hold a nest and from the total number of females visiting and remaining at the site for more than three days. Male success ultimately derives from the patterns of female associations with the males and nests. These associations of females to males and nests and their relationship to male success are examined in Chapters 4 and 5. The post-mating success of a male is defined as both rate of egg acquisition and guarding abilities. It will be measured as both the number of eggs a male acquired during his residency at a nest and as the total number of eggs hatched from the nests. Both the preand the post-mating

PAGE 35

27 success of a male are influenced by his expectation of further reproduction, and at any moment it is relative to the success of the other males in the population (Trivers 1972) . Since nests were used by different males, I have also investigated the success of the nests. This was done by determining the total number of eggs a nest acquired, the number of eggs hatched at the nest and the number of resident females associated with the nest. Nest success was estimated regardless of the identity or the number of nest owners. In the following sections, I describe the patterns of nest use and nest success. Second, I describe the relevant natural history of males, their behavior, and patterns of success. Third, I present a male switching experiment that separates the male and nest effects on breeding success. All correlations presented are Pearson correlations, unless otherwise stated. Thus I am assuming both a linear relationship between the variables and the normality of the variables (Sokal and Rohlf 1979) . The use of correlations for investigating the breeding success allows me only to establish associations between variables. They do not prove causation, which can only be established through experimental manipulations. The sample sizes for the tests reported will differ somewhat when considering different questions. This is because I do not have complete records on all aspects considered for all individuals and nests.

PAGE 36

28 Patterns of Nest Use Nest Re -use and Abandonment Not all nesting males guarded eggs in nests that they built. Nests often remained intact and were re-used. Acquisition of occupied nests through take-overs, and using abandoned nests were ways of obtaining previously-constructed nests . Use of nests from previous seasons. The permanence of nests throughout different breeding seasons was remarkable. Four out of the 172 nests monitored in 1985, were still present and were used in 1987. Three persisted until 1988. From Sites A and B, 21% of the nests (18 out of 86 nests marked in 1987) lasted one year. All nests remaining from the 1985 season that survived to 1987 were re-used at least for short periods of time in 1988. A total of 12% of the nests marked in 1987 (Sites A+B=86 nests) were re-used in 1988. Of the nine nests from Site A (1987) still present in 1988, four were used throughout the complete season, and six of the nine nests from Site B (1987) were occupied throughout the 1988 season. A total of 56% of the nests from 1987 lasting to the next season were reutilized in 1988. Take-ov er of nests . One way that males acquired an already-constructed nest was by means of a take-over. Take-

PAGE 37

29 overs involved fights between males (Mora 1987) . I attributed new ownership as the result of a take-over whenever I found a new male at a nest without any prior period during which the nest had been empty. Thus the takeovers that I report here were the between-census replacements of one male by another. This criterion may overestimate the rate of take-overs in the population while probably underestimating the rate of nest abandonment. Most of new male occupations at nests (63.5%, n=74, in 1987 and 56%, n=84, in 1988) were the result of nest takeovers (Table 3-4) . The frequency of take-overs was not different in 1987 from that of 1988 (X^=0.39, d.f.=l, p>0.05; Table 3-4) . Nest abandonment . Forty six percent of all nests monitored in this study (n=158) were abandoned by their resident males. Some males abandoned nests after they suffered damage from rain, ants or fungus (43%, n=72 nests abandoned) or for unknown reasons (57%, n=72) . Most males disappeared from nest sites (trees) after abandoning the nest (88%, n=72), but others remained in the same tree for the rest of the season (12%, n=72). A total of 37 (26.6%, n=158) new male occupations of nests in the 1987-1988 breeding seasons were on previously abandoned nests (Table 3-4) . Therefore, about half of the nests that were abandoned (49%, n=7 2) were never used again that year. Number of nest owners. Almost half the nests monitored in both seasons (44%, n=158) were occupied by different males

PAGE 38

30 throughout the breeding season (Table 3-5) . Multiple occupancy of nests was due to nest take-overs (36%, n=158) and use of previously abandoned nests (23%, n=158; Table 34) . Length of nest use. Nests were occupied by one or more males for an average of 2.5 months (Table 3-6). There was no difference in the length of nest occupation between years (t=-1.35, n2=59, n2=84, p=0.18, unpaired two-tailed t-Test) . Building New Nests Nests were made out of mud and tree bark collected by males from crevices of the trees and from deposits around the supporting structures of epiphytes. Nest construction behavior was previously described (Mora 1987) . Males compacted the nest material by rolling it with their chelicerae and pedipalps into tiny balls that they applied directly to the tree. The floor was laid out first and consisted of a circular, single layer of material and then, walls were raised vertically around the circular base. Even if a nest did not survive to the next breeding season, new males built nests at the same locations year after year. This occurred for four nests from the 1985 plot, three nests in the B Site, and two in the A Site. Four percent of the nests monitored during the two seasons (n=158) were rebuilt by their occupants at distances of 7-16 cm from their previous nest. At the forest sites, 82% of the nests

PAGE 39

31 marked in 1988 were new (80 out of 98) . Considering the availability of trees in this population, site fidelity was remarkable. These observations suggest that some trees, and particular locations on trees consistently attract reproductive males. Nest Measurements Nests had a mean height from the ground of 50.3 cm (SE=4.9, range 8-171 cm, n=98, Table 3-7). They had a mean diameter of 3.2 cm (SE=0.13, range 2-7.8, n=152. Table 3-7) and their walls had a mean height of 0.5 cm (SE=0.24, range 0.14-1.1, n:=152, Table 3-7). These dimensions were not different from those of nests sampled in 1985 (Mora 1987). Nest height above ground was not correlated with either nest diameter (r=-0.07, n=98, p=0.52) or the height of the walls (r=0.14, n=98, p=0.23). Also, the diameter of the nest was not correlated with the height of the walls (r=0.21, n=152, p=0.08). There were no differences between the two seasons in the height at which males built nests (Table 3-7). There were differences between years in the sizes of the nests: males constructed nests of greater diameter in 1987 than in 1988 (Table 3-7), and nests had higher walls in 1988 than in 1987 (Table 3-7). These differences were probably an artifact of the differences in quality of the calipers used to measure nests in the two breeding seasons.

PAGE 40

32 Patterns of Nest Success Is Egg Laving Random? Although some nests appeared to receive more eggs than others, this did not necessarily imply that oviposition was biased towards some nests. Were there some nests more likely to acquire eggs? Methods . To evaluate this question, I compared the distribution of ovipositions for a large sample of nests throughout a discrete period of time to that expected if the ovipositions were random. If the distribution of oviposition events was not random, more nests than expected would have no oviposition events and fewer visits (and then, "bad" nests do indeed occur) , and more nests than expected would have a larger number of oviposition visits (and thus significantly better than the unsuccessful ones) . There are several known distributions to which we can compare the observed distribution of ovipositions. Because many nests experience few or no visits (oviposition events were rare) , the Poisson distribution was appropriate (Sokal and Rohlf 1979) . I selected nests monitored in 1988 and evaluated the number of ovipositions experienced by 50 nests over a 30 d period. I analyzed data from September because that month had a large number of resident males present at nests for the 30 d period. I calculated the number of ovipositions as the

PAGE 41

33 ratio of the number of new eggs over the mean number of eggs per oviposition (x=3. Mora 1987). The distribution of ovipositions was compared to an expected Poisson distribution by a two-tailed Chi-squared test. Results . The distribution of oviposition events per nest was significantly different from the expected Poisson distribution (x=5.06, SE=0.78, X^=64.2, d.f.=3, p=0.001. Figure 3-1) . Discussion . Oviposition events were clumped. There were more nests with larger numbers of eggs than predicted by the Poisson distribution. There were some nests/males more successful at getting ovipositions; therefore, "good" and "bad" nests may in fact exist. However, the distribution of ovipositions is not bimodal . What may be the variables accounting for the differences among nests? What are the patterns of nest success in this population? These questions are considered in the following sections. Were Old Nests More Successful Than New Ones? Success in 1987. The nests that survived and were used in 1988 acquired more eggs than average in 1987 (for Sites A and B, t=3.67, ni=18 nests that persisted to the 1988 season, n2=20 nests that did not persisted, p==0.001, unpaired, twotailed t-Test) . However, nests that persisted and were used in 1988 showed no difference in hatching success in 1987 as

PAGE 42

34 compared with the nests that did not persist (t=0.74, ni=18, n2=20, p=0.46, unpaired, two-tailed t-Test) . Success in 1988 Old nests (18 nests from 1987 that survived and were reused in 1988 in Sites A and B) accumulated a mean of 29.6 eggs (SE=5.5, n=18) and hatched a mean of 61.4 juveniles (SE= 15.9, n=18). The number of eggs accumulated at these old nests was not different from the mean number of eggs at the new nests sampled that year in the C Site (Table 3-3, t=0.84, ni=18, n2=76, p=0.41, unpaired, two-tailed t-Test) . However, old nests in 1988 produced significantly more juveniles than new ones (mean juveniles for new nests= 27.7, t=2.11, ni=18, n2=76, p=0.05, unpaired, two-tailed t-Test) . The success of a nest in a year did not predict its success the following year. The number of eggs accumulated in the 18 nests monitored in 1987 that survived and were reused in 1988 was not correlated for the two years (r=0.16, n=18, p=0.6). Nor did the number of juveniles produced in 1987 correlate with those produced in 1988 in the same nests (r=0.4, n=18, p=0.18). Another measure of success, the association of females to those nests, will be discussed in Chapter 4 . In summary, nests from the 1987 season that persisted and were reutilized in the 1988 season, had been successful nests in 1987 in acquiring eggs. Those same nests in the following year showed a higher hatching success than new nests. The analyses showed that the success of any one

PAGE 43

35 individual nest in one year was not correlated with its success in the following season. Whether or not nests are consistently successful from year to year is not clear. From two seasons of observations, the pattern is that nests that are surviving and re-used were more successful than average (in terms of egg accumulation or hatching success) during both nesting seasons. However, the success of these nests was not consistently accounted for by either a higher number of eggs accumulated or by a higher hatching success. Was Len gth of Occupancv Associated with Nest Success? In both years, the length of nest use within a season was correlated with both the number of eggs accumulated at the nests and the number of juveniles hatched in them (Table 3-8) . These positive associations may have come about because more successful nests were occupied longer or because males' longer occupancy of nests increased the success of the nests. I have no way to distinguish between those two explanations . Are Take-over N ests More Successful? Nests that experienced take-overs held higher numbers of eggs (mean for nests that were taken over=37.9 eggs, SE=3.2, t=5.06, n=53, p=0.001, two-tailed t-Test, compare to averages in Table 3-12) and hatched higher numbers of juveniles than average nests in both seasons (mean for nests taken-

PAGE 44

over=39.5, SE=6.6, t=3.2, n=53, p=0.002, compare to averages in Table 3-12). These data suggest that male competition for nests, estimated by the rate of take-overs, was higher at more successful nests. Is Nest Success Correlated with the Number of Nest Owners? General correlations . There were no correlations between the number of owners a nest had throughout the breeding season and either the number of eggs that accumulated at the nest (r=0.08, n=101, p=0.41), or the number of juveniles produced at the nest (r^O.OOl, n=101, p=0.99). Since most new owners were due to take-overs (Table 3-4), this may seem contradictory to the observations indicating that nests experiencing take-overs were more successful than the average nests. Two reasons may account for the lack of correlations between the number of male residents and the success of the nests: (a) a male eats the eggs present at the nest after a take-over, decreasing the counts of eggs and juveniles accumulated at the nests and (b) new ownership of a nest may occur after a nest had been abandoned for some time. Therefore, many of the multiple occupancies occurred in nests that were empty for prolonged periods and thus had a lower accumulated number of eggs and juveniles for the season. Success of nests with more than three owners . Most nests were occupied by one or two different males, and a few

PAGE 45

37 had three or more resident males through the season. A comparison of the number of eggs and the number of juveniles between the 14 nests of this study that had three or more owners (Table 3-5) and a random sample of 14 nests with one or two owners (generated with a table of random numbers) showed the former had a significantly higher number of eggs (Mann-Whitney U Test, U=69, U crit=64, p<0.05), and hatched a higher number of juveniles (U=64, U crit=64, p<0.05; Figure 3-2) . These results indicate that males were indeed reusing more successful nests. As a consequence, the higher success of nests with more owners may be simply that there were more breeding attempts at those nests. I cannot distinguish if nests were more successful because they were reutilized more times and accumulated more eggs and juveniles, or if they were reused more times because they were indeed better nests than average . Are Aba ndoned Nests Less Successful? The success of abandoned nests was not different from the success of the average nest in the population. I have excluded the data from 1987 from this analysis because I do not have complete records for all nests abandoned in 1987. Abandoned nests in 1988 neither accumulated fewer eggs than the population mean (mean for abandoned nests=28.8, SE=4.2, t=1.77, n= 17, p==0.09, compared to the population mean in

PAGE 46

38 Table 3-12) nor hatched fewer juveniles (mean for abandoned nests=24.4, SE=6.1, t=0.97, n=17 , p=0.34, compared to population mean in Table 3-12) than continuously occupied nests. These observations suggest that males were not abandoning the less successful nests in the population. Is Nest Success Correlated with Nest Measurements? In general, the success of the nests was not strongly associated with most of the measured nest dimensions (Table 3-9). There was a weak correlation between the diameter of the nests and the number of eggs that nests accumulated, suggesting that nests of larger diameter may hold more eggs than smaller nests. The only nest measure associated with the success of the nests was the height of the nest's walls. Nests with higher walls hatched more juveniles than nests with lower walls (Table 3-9) . There are several reasons why walls may enhance the hatching success at the nests. Walls are physical obstacles that may deter the attack of predators such as ants and flatworms. Walls may also block water running down the tree after a storm and prevent eggs from being washed away. Such excess moisture might also increase fungal growth among the eggs .

PAGE 47

39 Male Life History and Behavior: General Descriptions Male Size Males had a mean weight of 0.014 g (SE=0.0002, range 0.008-0.2 g, n=227) . Their mean total body length, measured from ocularium to the posterior cephalothoracic spines, was 3.02 mm (SE=0.04, range 2.24-4.0 mm, n=227) and mean total body width, measured at the level of the fourth coxae was 2.5 mm (SE=0.03, range 1.6-3.12 mm, n=227). The mean length of the fourth tibia+patella was 3.67 mm (SE=0.01, range 3-4 mm, n=227) . This measurement most accurately describes the body size in opilionids (M. Goodnight, personal communication) and varied little in this population. Male Longevity Adult males can survive several years and reproduce for at least two seasons. From the 156 males marked at the A and B sites during the breeding season of 1987, four were recovered in 1988. Three of these had been occupying nests since the beginning of the season. The real proportion of individuals surviving to a second breeding season may be higher than this since the enamel I used for marking animals is not permanent. Individuals that survived to a second year can potentially reproduce throughout the complete breeding season. One take-over that occurred in November of 1988 was

PAGE 48

40 accomplished by a male marked mid-season in 1987, and one nest owner in his second year held a nest for the complete 1988 season. Length of Male Residency at Nests The average residency time of a male as a nest owner was 48 d for all sites over both seasons (Table 3-10) . This is only about a fourth of the total nesting season (JuneDecember) . Both the shortest and the longest residence times recorded for males were in nests at the Log site where there was a high concentration of nests on a single tree. This may have generated intense competition for nests and produced both very short and very long nest-tenures. Origin of New Residents Most (67%, n=94. Table 3-4) of the new owners of nests were not prior residents on the tree where the nests were located nor were they from trees adjacent to those nests. Rather, they were mostly unmarked males new to the nest site (Table 3-11 , x2=10.88, df=l, p< 0.001). It was easy to find males since I sampled the nest sites carefully, including the tree crevices, up to a height of 2 m. There is a possibility that the new nest owners had come from the tree canopy, but individuals in this population were not commonly found at heights over 2 m (Mora 1987).

PAGE 49

41 I suggested previously (Mora 1987) that non-guarding males stayed at nest sites probably waiting for an opportunity to take-over the nest or to use it after it was abandoned. Accordingly, I had expected that most of the new nest owners would be males marked at, or close to, the nest site. My observations however, showed that those males that had associated with the nest sites prior to the new occupancy were not necessarily the ones replacing the nest owner. Male movements were rather extensive, since the majority of new nest owners found between censuses were males not seen previously in the nest area. Nestless Males It is most intriguing that not all males built nests in the field. Most males built nests in captivity when a piece of bark was provided. I cannot estimate the proportion of nestless males since my observations were limited to nesting sites, but my impression is that it is large. I speculated (Mora 1987) that salivary secretions, which are added to the nest material at the moment of applying it to the tree (one can see the strings of silky material with the aid of a flashlight), may be metabolically costly or may develop only at a certain age. If the ability of males to build nests is constrained, this might account for the population of nestless males.

PAGE 50

42 In 1987, my assistant and I searched for the sources of nest building secretions by performing dissections of the cephalothoracic region of males that had finished constructing nests {n=4), of males collected at the moment of building a nest (n=2), and of wandering males not associated with nests (n=4). We found no evidence of a specialized gland. However, we did not utilize specific dyes, fixatives or equipment that may have enhanced our possibilities of finding such a gland, so its existence cannot be discounted. If they exist, the glands are probably very small and very fragile. Besides the repugnatorial glands located at the base of the first and second coxae, there are no accounts of accessory glands in opilionids (Berland 1949; Hillyard and Sankey 1989) . Male Behavior: Background The range of behavior patterns shown by Z. albomarainis is remarkable for an arachnid. The following account is of the activities of nesting males only. I have indicated with an asterisk (*) the behavior patterns for which descriptions were provided in Mora (1987). Each one of these behavioral patterns is a source of variability in breeding success of males. The relationship of some of them with male success will be evaluated in a following section. The different kinds of behavior patterns performed by males can be grouped into four general categories. (a)

PAGE 51

43 Foraging behavior . Males leave the nests to forage for food and nest materials. Males are omnivorous; they feed on insect larvae deposited in crevices, on carcasses of dead insects, on fruits and they hunt termites. They also leave the nests to gather mud and tree bark that they utilize for nest construction and repair. (b) Nest acquisition . Includes nest construction (*) , take-overs {* ) and searching for abandoned nests. The mechanisms and the mode of searching for nests and nest sites are not known. (c) Courtship and copulation . Male mate choice is important in this system since not all female courtship attempts lead to copulation and males chase off the nest some females that have initiated courtship (Chapter 4) . (d) Eaa guarding and nest maintenance . Guarding activities of males include nest cleaning!*), chasing off predators (*) , and repairing nest damage ( * ) . Components of Male Success Egg Acquisition Males guarded a mean of 21.4 eggs (SE=1.3; range 1-102, n=227; Table 3-12) during their residency time. Nesting periods, however, were not equal for all males. A better way to compare the success of males takes into account the residency times of males at nests. Correcting for the males' nest tenure, males acquired only one new egg every 2 d (x=0.5

PAGE 52

44 eggs/d, SE=0.06, n=227; Table 3-12). Both measurements of success underestimate the number of eggs acquired by males since many eggs are lost to predation (Mora 1987) . There was no difference in the number of eggs acquired by males over their residency time between seasons (t=1.87, ni=119, n2=108, p=0.06, two-tailed unpaired t-Test, Table 3-12). Hatching Success Nesting males hatched a mean of 18.5 juveniles over their residency time at nests (SE=2.3; range 0-211, n=227; Table 3-12). Taking into account the male residency time at nests and considering the total juvenile production per day, males hatched, on average, 0.3 eggs/day (Table 3-12). The 1988 season yielded higher nest production of juveniles than did the 1987 season (t=5.51, 1987=119 nests, 1988=108 nests, p<0.001, unpaired, two-tailed t-Test, Table 3-12) . The differences between sites in number of juveniles hatched were not significant (Table 3-12) except for the comparisons between the A and B Sites in 1987 (t=2.06, n=55, p=0.04, unpaired, two-tailed t-Test). Relationship Between F.a gs Acquired and Eaas Hatched The mean number of eggs at nests showed a positive correlation with the number of juveniles produced (r=0.47, n=201, p<0.001). There was a stronger correlation during the 1988 season (r=0.52, n=112, p<0.001) than for the 1987 season

PAGE 53

45 (r=0.46, n=89, p<0.001). The more eggs a male guarded during his tenure at a nest, the more juveniles hatched from the nest (Table 3-4, Figure 3-3). Some males appeared to have hatched more juveniles than the mean number of eggs they guarded (Figure 3-3). This result was due to the great variability of the data. Many eggs never hatched or disappeared from one census day to another. Because of those fluctuations in egg numbers at nests from census to census, I estimated egg acquisition from the mean number of eggs a male guarded during his residency at a nest. Since guarding males will only care for their own eggs (Mora 1987), juveniles hatched were most probably the males' offspring. Success and Male Size There were no significant correlations between the success of the males and any of their phenotypic characteristics. There was however, a tendency for large males to be slightly more successful than small males. There were overall positive but insignificant associations for both years between male weight and the number of eggs acquired (r=0.21, n=143, p=0.09), male weight and the number of juveniles produced (r=0.26, n=143, p=0.27) and the length of tibia+patella 4th and the number of juveniles hatched at the nest (r=0.36, n=143, p=0.07). Despite the fact that none of these associations was significant at the 0.05 level, they

PAGE 54

46 suggest that male size may have some influence on the success of nesting males. These are not surprising results. Body size has been recognized as contributing to male breeding success in many species (Clutton-Brock 1988) . In this species, large body size may confer a male advantage in competition for nests and in preventing predator attacks. Success and Mal e Lonaevitv The success of the four males marked in 1987 that reproduced again in 1988 was not different from the average success of males in those years. Those four "old" males acquired the same number of eggs and hatched the same number of juveniles as average males in 1987 (for eggs: t=0.04, n=4, p=0.96; for juveniles: t=-0.1, p=0.92, one sample t-Test) and in 1988 (for eggs: t=0.97, n=4, p=0.4; for juveniles: t=2.04, p=0.13, one sample t-Test). Males reproducing for a second season had higher success in the second than in the first year. The four males followed for two consecutive seasons guarded more eggs (z=1.6, p=0.05, one-tailed Wilcoxon Matched-pairs Signed-rank Test) and produced more juveniles (z=-1.83, p=0.03, onetailed Wilcoxon Matched-pairs Signed-rank Test) in 1988 than in 1987. This pattern of success is not uncommon. It has been recognized that several components of reproductive success improve with age in Drosonhila flies (Partridge

PAGE 55

47 1988), frogs (Howard 1988) and great tits (McCleery and Perrins 1988) among others. Since harvestmen do not grow after their last molt, part of this improved success may be derived from increased parental experience as occurs with great tits (McCleery and Perrins 1988) , or to their improved abilities to recognize and hold a good nest. Survival itself could be considered proof of a vigorous constitution, which may also result in good parenting and female preference for those males. Success and Male Residency Time The length of nest tenure influenced the success of males. Although the length of nest use was correlated with the number of eggs accumulated in the nest (Table 3-8), there was no correlation between the length of male residency at the nest and the mean number of eggs that the males acquired (r=0.13, n=101, p=0.19) during their nest tenure. Residency time was significantly correlated with the number of juveniles hatched (r=0.54, n=101, p< 0.001). These results suggest that a long residency time may not increase a male's chance to accumulate eggs, but it increases the number of juveniles hatched. Longer residencies may decrease the levels of predation by conspecific females. This could confer an advantage to long-term residents and may select for males holding onto nests for as long as they can. The intriguing question of why males abandon nests despite this

PAGE 56

48 strong association between residency time and hatching success will be addressed below. Success and Male Behavior It is possible that the guarding abilities or the mating success of males did not depend on size, but rather on their behavior at the nests. Are there any differences in nestrelated behaviors of successful as compared to less successful males? Method? . In 1987 season, my field assistant and I made continuous focal male observations at the Log site. I categorized males as "successful" when they had a hatching success of more than 20% (mean+ SD for all males at the Log Site) and at least one resident female. Males in nests with lower hatching success and no resident females were categorized as "unsuccessful". This categorization was done prior to the observations. We performed paired, simultaneous 3-h block observations on eight males, four in each category. The pairs were chosen using a random number table. The time of day for observation was randomized. Each nest was observed twice, for a total of 1440 min. figSUltg. Only successful males repaired walls even though both successful and unsuccessful males brought mud and bark to the nests (Table 3-14). Copulations were the only other activity exclusive to successful males. There was no difference in the number of times that successful and

PAGE 57

49 unsuccessful males engaged in nest cleaning activities (MannWhitney U Test for the amount of time spent cleaning, U=5.2, U crit.=16, p> 0.05), nor was there a difference in the amount of time males spent out of the nest (Mann-Whitney U Test, U=5.7, U crit=16, p>0.05. Table 3-14). Only unsuccessful males were observed chasing off other individuals . Discussion . Aggressive interaction was observed only in an unsuccessful male. This challenges my hypothesis that successful nests were competed for more intensely (as reflected by the higher rate of take-overs, which involve aggressive interactions) than unsuccessful ones. Aggressive interactions may diminish at nests after long residency or they may be too infrequent to be recorded in the time assigned for these observations. There were only two behaviors that distinguished successful from unsuccessful males: the occurrence of copulations and nest wall repair. I have shown before that the height of nest walls was the only nest characteristic associated with hatching success at the nests. It is not surprising then, to find that males that repaired walls more often were more successful than males that did not. I would predict that females may cue in on the condition of nest walls when assessing males. Manipulation of the height of the walls and its effect on the reproductive success of males could be an area of interest for future studies.

PAGE 58

50 Why do Males Abandon Nests? The possession of a nest is necessary for males to reproduce. Among the lines of evidence that suggest that there is competition for nests among males are the high rates of take-overs, the occurrence of nest reuse (Table 3-4) and the difference in success among nests (Tables 3-2, 3-9). I have also shown that the length of nest tenure is associated with hatching success at the nest. If nests are indeed difficult to acquire (there are nestless males in the population and there are take-overs and nest re-occupation after abandonment) and if longer tenure increases the hatching success, the question arises, why do males abandon nests? It is expected that the decision of whether to abandon a nest will be determined by the benefit of remaining at the nest as opposed to what a male will gain by leaving. The data one needs to evaluate male decisions on whether or not to abandon a nest are the fitness outcomes for (a) remaining in the nest and (b) alternative behaviors. We need to know the probabilities of a male acquiring another nest, the probabilities that the new nest will be better than the one he already has, the costs for searching for or taking-over a nest, the costs of keeping a nest for extended periods, the remaining life expectancy for these males, and how all the above change with the age of the individuals and throughout

PAGE 59

51 the breeding season. For this kind of analysis, I only have data on the take-over rate of nests. Another way to understand why males abandoned a nest site is to evaluate whether the males' probability of getting eggs changes over time. By doing this I am reversing the question and am now asking why males stay at nests. If the probability of acquiring eggs decreases with time we would expect males that have not acquired eggs to leave the nests. Conversely, if the probability of getting eggs increases with time, we expect males to keep nests for as long as they can. If the probability does not change with time, we expect males not to base their decisions of staying or not on the daily nest success. I have constructed a decay curve for the time it took a sample of males to get their first egg. The curve plots the log of the number of males without eggs versus time, following a group of males over 30 d and resembles the survivorship curves familiar to population ecologists. I have drawn curves for early-, midand late-season 30 d periods in 1988 (Figure 3-4). These curves showed a slow and relatively constant decay in the number of males with no eggs. The shape of the curves was the same throughout the season. They mean that there were about equal probabilities that each male will get eggs each day and that those probabilities did not change through the season. Consequently, the number of days a male remained without eggs should not influence his decision of whether or not to leave

PAGE 60

52 a nest. These results allow me to explain why males did not leave unsuccessful nests, but the question of why they did leave successful nests remains unanswered. One important reason that males left nests was the occurrence of nest damage. Nest damage was due mainly to ant and fungus attacks and heavy storms. The incidence of such damage and their possible effect on males' decisions to abandon nests will be discussed in Chapter 6. Nests or Males? A Male Switching Experiment I have been interested not only in identifying the extent of the variation in male reproductive success, but also the phenotypic and ecological variables that account for the observed differences. Such information is necessary to evaluate the operation of natural and sexual selection and the evolution of morphological and behavioral adaptations. In this species, the main problem with addressing the question of what variables may be important for explaining variation in male success is separating nest and male effects since male success is dependent on the possession and retention of a nest. Addressing this problem is also important because nest success also reflects the reproductive success of the females that oviposit in them (Chapter 4).

PAGE 61

53 I attempted to separate the effect of the nest from the success of males by performing a male switching experiment. I switched both successful and unsuccessful males to previously successful and unsuccessful sites and compared their success, in terms of the number of resident females retained, before and after the switching. If male success is mainly determined by the nest, the success of an unsuccessful male will increase when switched to a successful nest. If females are associated with the nest rather than with the males, they would not leave and would continue ovipositing in a nest regardless of what male was present. The specific predictions were that if the success of males prior to the switching was largely due to the nest, it would decrease when switched from a good to a bad nest. If the success of males was not strongly associated with the nests, the success of males would not, on average, change when switched to a new nest. Methods . I considered successful males as those with more than one resident female (a female associated with the nest for more than 9 d, i.£. , that will potentially oviposit twice), and with a hatching success 20% or more, i.^. , above the mean for the population at mid-breeding season. Bad and good nests were defined using the same criteria. In September of 1988 (mid-breeding season) , considering nest sites with only one nest per tree, I was able to identify a total of 15 successful males at good sites and 18 unsuccessful males at bad sites. I randomly assigned males

PAGE 62

54 to new locations. Seven successful males were switched to good sites and eight were switched to bad sites. Of the 18 unsuccessful sites and males identified, eight males were switched to previously successful sites and 10 were switched to unsuccessful sites. Only one pair of males was exchanged through this random assignment. By randomization, one individual was to remain in its original site (one time in this experiment). I gently picked it up and removed it for 4 h before returning it to its original nest. I measured the success of males before and after the switching by the number of females that remained at the nest sites because males fed on the eggs present at the nest after their arrival at a new nest, and because it takes a variable and long time for the males to accumulate and hatch eggs in a new nest. Data on the number of females resident on trees before and after the switchings for each nest, were compared using Wilcoxon Signed-Rank Test, one-tailed, at an alpha level of 0.05. Results . When males were switched, successful males decreased their success, in terms of number of females potentially ovipositing at the nests, when switched to bad sites (Wilcoxon Signed-Rank Test, t=0, n=10, p<0.001, onetailed) . Conversely, previously unsuccessful males experienced an increase in the number of females potentially ovipositing at their nests when switched to good sites (Wilcoxon Signed-Rank Test, t=0, n=8, p<0.05, one-tailed. Figure 3-5) .

PAGE 63

55 Discussion . These results suggest that the success of males depends on the nest/nest site. A male can increase his success by associating with a good nest. The quality of nests may be the most important factor accounting for the variance in male reproductive success. Discussion: Patterns of Nest and Male Success I have utilized three measures of success to discuss the patterns of nests and males success. These measures are the male occupation of nests (reflected as length of nest use when discussing nests or tenure of nests when discussing males) , the number of eggs acquired and the total number of juveniles produced. Additional measures of nest and male success can be derived from the association of females to nests/males and will be discussed in the next chapter. The most important pattern emerging was that nest site was the most important component of the success of nests and males. There are four general assumptions implicit in my evaluation of the variation in male success. (a) A resident male fertilized all the eggs he was guarding. This assumption is supported by my observations that males fed on eggs that were not their own and they invariable copulate prior to oviposition (Mora 1987). (b) Eggs accumulated at nests were oviposited by the resident females. The number of

PAGE 64

56 females visiting nests between censuses is unknovm and their contribution to nests is not considered in this study. The validity of this assumption is evaluated in Chapter 4. (c) Males with no nests had no success. I have not observed females mating or ovipositing outside a nest. (d) Individuals and nests marked and followed were a random sample of the total population of nesting males. Patterns of nest use . The distribution of nests on the island was patchy. Not only were nests concentrated in the central portion of the island (Figure 2-1), but within the study areas, males built nests in the same locations or a few centimeters away from former nests year after year. There was a high rate of nest reuse; nest take-overs and the utilization of abandoned nests resulted in the multiple occupancy of nests. Half of the nests monitored were abandoned at least once during a breeding season and more than half of the nests experienced at least one take-over (Table 3-4). There were practically no nests that stayed the same for a complete season. These observations also suggest that there is something about a nest or the nesting site that attracts males and results in the reuse of some nests. Patterns of nest success. The number of ovipositions among nests was not random. More nests received very few and more nests received more visits than expected if ovipositions were random (Figure 3-1). This suggests that there were indeed "bad" and "good" nests in the population.

PAGE 65

57 Nest longevity was a source of variation in nest success. Nests from the previous year (marked in 1987) were more successful in 1988 than nests constructed that year. The length of nest occupation also produced differences in nest success. The longer a nest was occupied, the more juveniles hatched in the nest (Table 3-8) . Two observations suggest that "good" nests experienced higher levels of competition among males: nests that suffered take-overs were more successful than nests that did not, and nests that experienced multiple occupancy, i.e., with three or more nest owners, were more successful than nests that were not taken over or had only one or two different male residents (Figure 3-2) . Pattern s of male behavior and success . The variation in male success was high when measured both as the number of eggs males guarded (Table 3-12) and as the number of juveniles hatched at the nests (Table 3-12). Males traveled great distances searching for nests; most new owners were new to the nesting site (Table 3-11) . Surprisingly, I found few nest-related behavioral differences among successful and unsuccessful males. Apart from copulations, the single behavior at the nest observed only in successful males was repairing nest walls (Table 3-14) . Since height of nest walls was the only characteristic of the nests correlated with nest success (Table 3-9), the amount of time and the frequency that males repaired walls may influence male success. A male's ability to repair walls may

PAGE 66

58 be related to his age or breeding experience. Repairing a nest wall implies the same process as constructing a nest; both require the addition of salivary secretions to compact and cement the nest material. The availability of such secretions may be functions of age or physiological condition. Dissections of males did not disprove that there may be a constraint on nest construction. Reutilization of nests suggests that nest construction may be costly for males. The association between height of walls and nest success may be due to higher walls preventing egg loss, from females choosing to oviposit in nests with high walls because they reflect parental abilities or the genetic quality of the male, or from both. The ability of a male to maintain the nest walls may also be affected by the quality of the nesting site. Sites likely to receive large amounts of rain running down the tree trunks or exposed to the attack of predators may be harder to maintain than safe sites. It was surprising to find no strong association between size and success of males. Body size has been identified as an important component of the male's reproductive success in most species as larger size confers competitive advantage to males (Clutton-Brock 1988). Males of Z. albomaraini s do not follow this pattern. This is not surprising for an arthropod. Males of most species of insects are smaller than females and rely on other visual, acoustic or chemical signals to compete with each other. Also, in species with high paternal investment other phenotypic and ecological

PAGE 67

59 attributes affect breeding success. In many birds in which the male contributes to the caring of the young, nest site quality and breeding experience are the two main factors affecting the male's breeding success and not body size (Arak 1983) . Males can survive through and reproduce for at least two seasons. This means that older, more experienced males are part of the breeding population, together with young males in their first breeding attempts. Males that reproduced a second year had higher success than the average males in that year. Although the sample size was small and there was no evidence of behavioral differences among old and young males, these observations suggest that age was an important factor affecting the breeding success of males in this species. One possible manner in which age is related to male success is that old males are experienced and have proven survival attributes. The same physical attributes that permitted longevity may be important for increased reproductive success . Is is unclear why some adult males in the population do not build their own nests. The size of the population of nestless males is unknown but is presumed to be extensive; they apparently have no reproductive success. The males' tenure at the nest site was associated with their success; males that stayed the longest had higher success (or vice versa) . Since the probability of a male getting his first egg at a given site does not change with time during the

PAGE 68

60 season (Figure 3-4) , males should be selected to keep a nest for as long as they can. Why males abandon nests is not known. The success of the nests abandoned was not different from the average nests in the population. The extent to which nest damage may have caused nest abandonment will be discussed in Chapter 6. The mal e or the nest? There are several reasons that suggest the nest (site) is an important factor affecting male success. The high rate of nest re-utilization and the fact that not all males built a nest suggests that nests are expensive for males. The use of the same spots for two seasons suggests that not all sites are suitable nesting sites within the study area. The male switching experiment showed that females stay with nests after a male switching. Thus, the success of males was directly affected by the quality of the nest/nesting site. A clear pattern emerges that variation in quality of nests and nesting sites may be the most important factor contributing to the variance in male reproductive success. The strong competition for particular nest sites, evidenced, among other ways, in the high turnover of nest owners may suggest that sites themselves are important in male decisions to challenge an owner and female decisions to stay at a particular nest site. The importance of the site for a female's decision to remain associated with a particular nest will be discussed in Chapter 5; the factors that may account for differences among sites will be discussed in Chapter 6.

PAGE 69

61 Table 3-1. The number of nests marked and monitored for the 1987 and 1988 breeding seasons. No. of nests Total No. Year Site monitored nests marked 1987 Log 36 39 A 22 56 B 16 30 Totals 74 125 1988 A 11 11 B 15 15 C 58 72 Totals 84 98

PAGE 70

62 Table 3-2. The number of individuals marked and followed, bysite and sex for the 1987 and 1988 breeding seasons. Individuals Followed Individuals Marked Year Site Males Females Males Females 1987 Log 63 102 79 >220* A 40 Z J D 57 361 B 16 38 20 45 Totals 119 376 156 >626 1988 A 12 36 16 66 B 20 73 38 120 C 76 319 183 470 Totals 108 428 237 656 Overall Totals 227 804 393 1282 * I marked approximately 40 more females at this site .

PAGE 71

63 Table 3-3. Nest densities on buttressed and unbuttressed trees for the 1987 and 1988 field seasons combined. Number of nests on tree Tree Tru nk Tvne 1 2 >3 Buttressed 60 11 34 Unbuttressed 109 7 2 42.9, df=2, p<0.001

PAGE 72

64 Table 3-4. Nest re-use for the 1987 and 1988 breeding seasons . Mode of New Utilization Used after Ygar SitQ Take-ove r Abandonment n 1987 Log 11 17 36 A 16 3 22 BO 0 16 Total 27 20 74 1988 A 4 4 11 B 3 2 15 C 23 11 58 Total 30 17 84 Overall Totals 57 37 1^

PAGE 73

65 Table 3-5. Number of males occupying the nests monitored during the 1987 and 1988 breeding seasons. Number of Males Occupying Nest Year Site n 1 2 3 >4 1987 Log 36 20 11 3 2 A 22 7 12 2 1 B 16 16 0 0 0 Totals 74 43 23 5 3 1988 A 11 7 2 1 1 B 15 10 5 0 0 C 58 28 26 4 0 Totals 84 45 33 5 1 Overall Totals 158 88 56 10 4

PAGE 74

66 Table 3-6. Total time (in days) nests were occupied by one or more males during the 1987 and 1988 breeding seasons. X£a£ Length of nest use-SE (d) range 1987 82.4 5.27 8-161 59 1988 • 75.8 ± 4.85 6-160 84 Qygrall 78.7 ^ 3.57 6-161 141

PAGE 75

67 0) •H aJ T3 0) 0) G O 0) M aJ c 6 T! T) 0) i-i O JJ -H O e o u o ^1 (D 4-1 0) in m 0) 0) I O 4J W JJ J-) -H m M 4-1 ^ o o u U t3 C G (0 w CO +1 0) > m • 4-) 00 C 00 e 1 CO CO c m (^ 00 i^D CO ^ — 1 t~H O O O PI O * O o O V in ro • m o • 1 1 m £ o CO rH o U r~1,0 o O o o W +1 + 1 +1 + 1 +1 +1 w +1 rH CTl in -i 14H 4-1 ^ O
PAGE 76

Table 3-8. The correlation between the length of nest use (d) and the number of eggs accumulated at the nests and the number of juveniles hatched (Pearson correlation coefficients) . Year Variable r n P 1987 Eggs 0.35 59 0.008 Juv. 0.39 59 0.004 1988 Eggs 0.29 84 0.01 Juv. 0.66 84 <0.001 Overall Eggs 0.29 143 0.006 Juv. 0.45 143 0,001

PAGE 77

69 Table 3-9. The correlation between the nest measurements and the number of eggs accumulated in the nest throughout the season (Eggs) and the number of eggs hatched (Juv.) at the nests. (Pearson correlation coefficients) . Nest Variahlp n Eaas P Juv. 1? Height from ground 98 0.07 0.66 0.12 0. 48 Diameter 152 0.30 0.06 0.07 0. 68 Height of walls 152 0.10 0.50 0.35 0. 02

PAGE 78

70 Table 3-10. Mean residency times (d) of nest owners for the 1987 and 1988 breeding seasons. Year Site Mean davs (SE) ranae n 1987 Log 49. 1± 5.03 3-150 63 A 53. 9± 5.05 5-145 40 B 15.61.58 4-17 16 1988 A 59.6^11.61 15-119 11 B 56.17.19 7-122 20 c 53. 5^ 4.10 6-142 80

PAGE 79

71 o 4J ^! Cn C •H u (U (0 CQ iJ CO 0) c •H 0) ^: o M-l CQ 0) CQ 5 •H Cn •H O I CO 0) Eh CQ CQ C O CQ m CO cn G •H TJ 0) 0) 4.) C 0) Tl •H CQ
PAGE 80

72 r~ 00 CO 00 a\ o iH iH E-t CN O D • 1 • • 1 • • 1 • • 1 iH < tH CN O o o o ^ O CD o o o CN n o^ o in • ^ rH • d • CN rH 00 O rH CTl CN 00 O O CN • • 1 o rH • 1 • • 1 • • 1 CN cn W T3 Id \ (U c: d) -H i» C/3 S w K u w cc; S W OS >-) U) ct, M cn cn cn (U CU rH •H (0 d e e > u II o r-i T3 cn 4H 0) J-) O ^3 cn U (D u tC C 0) :3 X! e 3

PAGE 81

73 Table 3-13. Correlation between the number of eggs males acquired and the number of juveniles produced for the two breeding seasons studied (Pearson correlation coefficients) . Year Site r P 1987 Log 0 .67 <0.001 63 A 0 .40 0.01 40 B 0 .87 0.001 16 1988 A 0 .77 0.006 12 B 0 .23 0.24 20 C 0 .58 0.0001 76 Overall Totals 0 .47 <0.001 227

PAGE 82

74 Table 3-14. Behavior patterns shown (total number of occurrences, except for time out of nest) by successful and unsuccessful males during 48 h of accumulated focal animal observations at the Log Site. Each male was observed for a total of 6 h in two, 3-h blocks. Type of Male Activity Successful Unsuccessful (n=4) (n=4) Out of nest 72.3-30.4 47.0-18.7 (x SE min) Brings materials 1 1 to nests Cleans nest 3 1 Chases other individuals 0 1 Repairs walls 3 0 Copulations I Q_

PAGE 83

75 14 12 a P 00 10 0) 8 4-1 0 6 No. 4 2 0 Expected Oviposit ions 0 Observed Oviposit ions [ a 11 ,n — J3. Id OHa>(7\or-ir«o^tr>vor> No. Of Ovipositions/ 30 d Figure 3-1. Frequency of ovipositions (No. of new eggs/3) observed over 30 d in 50 nests and the expected Poisson frequencies .

PAGE 84

76 a s 50 40 30 20 10 0 O Nests with less than 3 owners (n=14) Q Nests with 3 or more owners (n=14) X I Eggs Juveniles Figure 3-2. Number of eggs and juveniles in nests that had three or more resident males through the 1988 season compared to nests that had one or two owners. Sample nests were derived from the C Site.

PAGE 85

4 11 0) A u n > o 200 160 1 120 80 -I 40 0 J 20 40 60 80 100 No. Eggs Accumulated Figure 3-3. Number of juveniles produced at a nest as a function of the number of eggs accumulated in the nest throughout the male's residency, for both seasons studied (r=0.47, n=201, p<0.01) .

PAGE 86

78 Figure 3-4. The probability of a male getting his first egg. Males were followed over 30 d during early, mid and late season of 1988.

PAGE 87

79 O rH (d 6 a V iH (0 P ide ith m 0) (U No iat o a 0 (0 a 0) s •< 4 n 3 2 1 8 Before Switching After Switching X 10 G/G G/B B/G B/B Male and Nest Site Figure 3-5. Changes in male success, measured as the mean number of resident females associated with males at good and bad nest sites after the switching of previously successful and unsuccessful males. Success of males was monitored 15 d before and 15 d after the switching. Sample sizes are indicated above the columns. G/G=males from good sites switched to good sites, G/B=males from good sites switched to bad sites, B/G=males from bad sites switched to good sites, B/B=males from bad sites switched to bad sites.

PAGE 88

CHAPTER 4 FEMALE ASSOCIATIONS WITH NESTS AND FEMALE SUCCESS The reproductive success of both females and males is influenced by where females oviposit. How many and which nests to visit, whether to eat eggs or not, whether or not to court and copulate with a particular male or to leave a nest, and, if they stay in a nest site, for how long, are some of the decisions involved in determining which nests females use. There are different trade-offs associated with each of these decisions (Dunbar 1983). The decision rules may change according to the interaction of females with males, the physiological condition of the female, and ecological variables related to the nests and the nest sites. Understanding female mating decisions is crucial for understanding the origin of the mating system (Wittenberger 1983). The patterns of female decisions are based on the relative value of material and genetic benefits and the correlation in the quality of benefits a male may offer (Borgia 1979). The importance of female versus male mate choice and what are the criteria for choice are interesting for species with male parental care (Thornhill 1979). They may provide insight into the operation of sexual selection in a particular system. This chapter examines the association of females with males/nests and explores the relationships 80

PAGE 89

81 between female behavior and the reproductive success of females and males. The study of female behavior was difficult. When trying to understand the behavior of females I found it was easier to follow and quantify the behavior of females who remained at nest sites than the behavior of those that did not. For isolated nests, the assumption that females remaining at the nest site were laying eggs at that nest appeared reasonable. I assigned eggs accumulated in the nest to the resident female(s). For nest sites with more than one nest per tree, I was not completely sure which female was ovipositing where, unless I directly observed the copulation and oviposition events. Consequently, I have limited most data analyses of female behavior to nest sites with one nest per tree. Some analyses are presented however from the Log study site since it permits me to compare the behavior of females under the same nest and male availability and general site conditions. Types of Visits an d Female Condition Females of Z.albomarainis mature a few eggs at a time throughout the breeding season. When ready to lay eggs, they enter nests, initiate courtship, oviposit and leave the male guarding the eggs (Mora 1987). They wander from tree to tree visiting nests or they remain associated with nest sites for variable lengths of time.

PAGE 90

82 Females may visit nests for different reasons: to mate, to assess the male and the nest, or to feed on eggs. I have categorized and described in a previous study female nest visitation in the following way (Mora 1987) : (a) in-and-out visits (03.. 70%, n=137 visits observed in 1985), (b) courtship (successful or not) and copulation visits (28%, n=137), and (c) visits leading to male chasing the female off the nest (2%, n=137) . One reason for these differences in female behavior and male chasing of some females may be differences in the reproductive condition of females. If correct, I would expect that females chased from nests would have fewer mature eggs than those that successfully courted a male. These rejected females entered nests only to feed on eggs. Or it may have been that males could assess female fecundity and rejected courtship from females not likely to oviposit in his nest right after copulation. This explanation, however, does not have clear predictions either for the function or for the reproductive condition of females performing in-and-out visits. For example, females assessing the males and nests might or might not have been ready to lay eggs. Did males chase off and bite females of lower than average fecundity? What was the reproductive status of females visiting nests and did it differ for females performing different kinds of visits?

PAGE 91

83 Methods During the breeding season of 1987, I accumulated with the help of my field assistant, more than 120 h of focalfemale observations. We performed the observations opportunistically, selecting nest sites with good visibility and with guarding males at the nests. We collected females that were carrying out different kinds of behaviors. We dissected those females in the laboratory for the number of mature eggs, that is, eggs that were the size and color of those being oviposited. Females were divided into five behavioral categories: (a) Females remaining inside the nest for more than one hour (n=13) . Females entered the nest and stayed motionless for prolonged periods of time, which were always more than one hour and extended for up to three hours, the maximum time allocated to the observation of a single female. I considered this behavior part of the courtship sequence in a previous study (Mora 1990) . (b) Females leaving the nest after ovipositing (n=8) . These females were observed ovipositing, and were collected as they were leaving the nest. (c) Females going in-and-out of nests (n=7) . These females entered a nest, tapped the substrate and left. (d) Females chased off the nests and/or bitten by the male (n=6) . These females were collected as the male was running after them. Two of these females were bitten after laying one egg.

PAGE 92

84 (e) Wandering females (n=12) . These were females collected from trees that had no visible nests. Results Three kinds of eggs were found in the dissected females. Large numbers of small white eggs were found close to ovaries; their numbers were too large to count accurately, but varied between 6 and >200. A small number of medium sized yellowish eggs were found along the oviduct; these were not found in all females and their numbers never exceeded six. Large yellow eggs were found at the distal portion of the oviduct and filled most of the abdominal cavity. We considered these mature eggs ready for fertilization since they were the size and color observed during oviposit ion. The following analysis refers to the latter category of eggs. Females from different behavioral categories had different numbers of mature eggs (Kruskal-Wallis one-way Anova, H=15.75, d.f.=4, p<0.05, Figure 4-1). There was a gradation of the mean number of eggs ready for oviposition within the females dissected. At one extreme, the females remaining inside a nest for extended periods had the most eggs available for fertilization. On the other end, wandering females had the fewest . Females leaving the nests after laying eggs in them had the second highest number of eggs ready for oviposition, followed by females bitten or chased off the nests. Females performing in-and-out visits

PAGE 93

85 had the least number of mature eggs of all females visiting nests (Figure 4-1) . Discussion These data showed that the physiological condition of females influences the kinds of visits they perform. They support the assumption that females arriving at nest sites, entering nests, and staying at nest sites were all females which potentially could copulate and oviposit. Those wandering on trees with no nests were not contributing eggs to the population at that moment. I predicted that chased off females were those entering nests only to eat eggs and therefore would have fewer eggs than those not chased. This was not true. Females that finished laying eggs still had a high number of mature eggs and this may explain why females remained associated with nests for prolonged periods. This observation also supports the assumption that those females that remained at the nest sites for more than three days probably oviposited at least once in that nest. Females sitting at nests were probably waiting to copulate and oviposit whereas wandering females were or may have been ripening eggs rather than looking for oviposition sites. The latter point raises the interesting question of why females may choose to ripen eggs away from nest sites. It is difficult to interpret the relationship between female nest visitation behavior and reproductive condition.

PAGE 94

86 One cannot predict the kind of visit a female may perform to a nest from her reproductive condition, except for the females with very few or very many eggs (at the extremes of the distribution) . One male behavior related to the female reproductive condition is intriguing. Why did males chase off females with ripe eggs? The possible explanations for such behavior are that (a) males chased off females that visited the nests to feed on eggs, (b) males rejected females because additional eggs may be detrimental to those already present in the nest and (c) males rejected females on bases other than their fecundity at the time of the visit. I do not have information on the previous association of the rejected females to the nests from which they were chased off, but I would expect that females that have not oviposited in the nest would be more likely to feed on the eggs present than resident females would. Male mate rejection of unfamiliar females may explain the association patterns of females with nests discussed in the following sections. Since the hatching success of nests was correlated with the number of eggs a male gets to guard (Figure 3-3, Table 3-13), there does not seem to be any reason for guarding males to avoid getting additional eggs. Hence, the detrimental effects of receiving additional eggs may not be the reason for males to reject females. In species where paternal care occurs, male mate choice is expected to become more important than for species in which males provide low parental investment

PAGE 95

87 (Ridley 1978; Searcy 1982). Male mate choice could be based on a number of traits not measured in this study that may indicate the genetic quality of the females (Petrie 1983) . Whether or not Z.albomarainis males are choosing those females with "better genes" is a question that will require laboratory mate choice experiments. Patterns of Female Asso ciations Operati onal Sex Ratios I estimated the sex ratio in the nesting population to be 1 male:3 females (393/1282) . There was a greater proportion of females in relation to males in the 1987 than in the 1988 nesting season (Table 4-1) . The proportion of males to females varied between study sites but these differences were not consistent between years. The sex ratio of nesting adults was most biased in the A Site, with a proportion of six females for every male. These differences in sex ratios between sites suggest a patchy distribution of individuals. Consequently, not all males face the same level of competition to attract females and the competition for nests among females may vary from one patch in the forest to another. Female longevity . Females can survive to a second breeding season; three females from 1987 were observed visiting nests in 1988.

PAGE 96

88 Frecruen cv and Patterns of Visits Most females observed around nests had eggs ready for fertilization and oviposition and females had the capacity to travel from tree to tree in search of males and nests. Since nests with guarding males were abundant in the study area, one may assume that females arriving at nest sites were looking for oviposition sites and therefore were faced with a series of decisions: (a) how many nests to visit, (b) which nests to visit, (c) stay or leave a nest site, and (d) if they stay, for how long. Following, I will describe the overall population patterns of nest visitation. Methods . The best way to quantify the number of nests visited by females would have been to follow the movements of marked individuals through the entire study area for extended periods of time. I have found this practically impossible at the forest sites. Once a female left a marked tree, the chances of seeing her again were very low. Also, most of the movements were carried out at night when observations were difficult. As such, I chose to estimate the extent of female movements between nests from the rates of female immigration to nest sites considering only trees that had one nest. These movements were recorded as part of the regular population censuses. This estimate is conservative because it is based on only one observation every 2 d per site, and

PAGE 97

89 it does not consider movements between trees with multiple nests . I also analyzed female visits to nests for 48 h of accumulated observations in 3 -hour blocks at the Log Site. Here one may assume that analyzing the female movements from nest to nest may give an accurate estimate of the female tendency to visit several nests, controlling for the associated decision of whether to stay on the tree or not. Following female movements within the same tree has the added advantage that i could observe several females simultaneously under the same conditions of time of day, nest availability and meteorological conditions. Results . Female movements were remarkable in the forest: marked individuals were recorded traveling more than 500 m from the nest at which they were marked over a two-day period. The mean rate of female arrival at nest sites was 0.52 female/day (SE=0.27, range=0-3.5, n=145). The mean total number of different females that visited an individual nest site was 5.11 (SE=0.77, range=0-23 females, n=145) for all forest sites. Twenty-two percent of all nest sites never received any female visits and less than four percent received visits from more than 20 different females during the breeding season (Figure 4-2) . At the Log Site, individual females visited a mean of 0.87 nest/day (1.73 nests over 48 h in 3-h blocks, SE=0.18, n=77); 25% visited none, and one female visited six nests. The frequency of visits to nests at the Log Site was not I !

PAGE 98

90 significantly different from that at the forest sites (X^=0.18, d.f .=1, p>0.05) . Discussion . Under the same conditions of nest availability and in the same nesting area (as in the Log Site) , females showed individual differences in mobility between nests. However, the extent and consistency of these differences was not examined. This pattern of nest visitation may have been the result of different rates of movement between females or by the different ability of males/nests to attract visits. Individual differences in mobility may also be present in females visiting nests at the forest sites. Forest females, however, must travel greater distances than females at sites with multiple nests. The cost of visiting nests may vary for females on trees with multiple nests as compared to females on trees with a single nest. It would be desirable to measure experimentally the differential survivorship for females in a log and a forest site or to measure the time between ovipositions for females that visit several nests in the same tree as compared with females who visited several trees. The difficulty of following females from tree to tree at the forest sites made such analyses practically impossible . There were nests for which I had no recorded female visits for the entire breeding season and nests that attracted a large number of females. Is this a random

PAGE 99

91 pattern or are females attracted differentially to certain nests? Do Females Visit at Random? Some nests in the forest site received no female visits during the nesting season while others received up to 23 visits (Figure 4-2) . This variation is important because it suggests the existence of female preferences. Such a pattern of female visits may have been generated because females are indeed attracted to some nests more than others, or it might have been generated randomly. Are there in fact nests receiving too many or too few visits as compared to the visitation frequency expected by chance alone? Methods . I randomly selected from the forest sites 50 nests occurring singly at trees. I extracted from the census data the number of different females that visited the nests for a 30 d period in the 1988 season. I considered a visit the appearance of an unmarked female or of a marked female new to the tree. I calculated the distribution of the number of visits per nest and compared that to an expected Poisson distribution. The two distributions were compared with a two-tailed, Chi-squared test with an alpha level of 0.05. Results . The mean number of females that visited a nest over a 30 d period for 50 nests was 2.6 females (SE=1.26, range 0-9, n=129. Figure 4-3). The distribution of female

PAGE 100

92 visits was not different from a Poisson distribution (x2=6.07, d.f.=5, p=0.29. Figure 4-3). Discussion . If females arrived on trees in a non-random pattern, I expected to find too many nests with no visits and more nests than expected with a large number of female visits. The results indicate females visited nests randomly and the proportions of nests with zero or with any number of visits were not different from those expected by chance. Although this analysis refers only to trees with one nest, the result was surprising since it shows that females were not attracted to certain nests more than others. Female movements did not follow directional or biased patterns that would have produced clumping of the visits. A random pattern of nest visitation may be generated by random encounters of females with trees. This may partially explain why males remained at nest sites even when they had received few or no female visits (Chapter 3). Since arrival of females at trees is random, a better estimate of female choice may be derived from the female residency time at nest sites and from their rates of oviposition than from their rates of visits to nest sites. Length of Female Associa tions to ^Jp<:l^s Female Residency. i assume that females arriving at a nest site could remain at a given nest site or leave to search for another. The possible association of females to

PAGE 101

93 nests ranges from females that were recorded at a particular nest site only once during the breeding season, to females that never left the site. At most nest sites, there was a dynamic and ever-changing association between a resident male and visiting females. For others, the result was a long term association between a resident male and resident female (s). In these situations there was a variable number of females overlapping at a nest site. I considered a female resident at a nest site if she remained at the site for more than 3d. I had calculated that the average time between oviposit ions was 3 d (Mora 1987) . Females at sites for at least 3 d would potentially oviposit at the nest once. I recognize however, that a female there for 1 d could oviposit if she had been wandering several days. About half of the females arriving at nest sites remained at the nest site for more than 3 d (mean number of females staying at sites for more than three days=2.59, SE=0.45, range=0-ll, n=145; Figure 4-4). The number of resident females was significantly associated with the number of females visiting the nesting site. The total number of different females that visited a nest through the nesting season was correlated with the number of females remaining at a nest site for more than 3 d (r=0.82, n=145, p<0.001; Figure 4-5) . Considering the longest period spent at each nest by a resident female, the mean longest residency time of females at the forest nest sites was 40.5 d (SE=6.62, range=3-160 d.

PAGE 102

94 n=143 females; Figure 4-6) . Mean maximum residency time at the Log Site was 43.3 d (SE=4.72, range=l-150 d, n=77 females) . Female overlapping at nest sites. Females overlapped at nest sites. These associations were variable. The mean number of females occurring simultaneously in a nest site at a given time was 2.33 females (SE=0.17, range=0-8 females at site, n=142) . Female-Female Interactions The overlapping of females at sites was sometimes associated with aggressive female-female interactions. I had previously observed nests for prolonged periods and noticed that some females associated with particular nest sites were aggressive towards other females that came close to those nests (Mora 1987) . These interactions consisted of females running after one another, and, very rarely, biting each other. These aggressive interactions were rare. I observed them only nine times at the forest sites. On two of these occasions I could not determine which female initiated the interaction. For the remaining seven cases the females that initiated the aggression were resident at the nest and directed the aggression towards newly arrived females (n=5) , a female with shorter residency at the nest {n=l) or a female of longer residency at the nest (n=l) . The success of the

PAGE 103

95 nests in which aggressive female-female interactions were observed was not different f rom the average success of nests in the population (sample x eggs=19.2, t=1.63, p=0.10; sample X juveniles=20 . 7 , t=1.68, p=0.09; two-tailed t-Tests) . Females may be able to exclude other females from the nest sites, or establish dominance over females of shorter residency at the nests through aggressive interactions. If males choose mates based on familiarity, new resident females may increase their chances of obtaining mates by extending their residency at a nest site. Resident females of longer tenure may decrease their chances of obtaining copulations with the addition of each new female. There are other species in which the length of female associations with a male or a territory (in the form of age or residency time at a territory) establishes dominance hierarchies (Baylis 1981; Clutton-Brock 1982). Future studies should be directed to test whether females of longer residency monopolize most matings . Female Nest Guarding Occasionally females remained "guarding" the nests after the male left. Some males left the nests only for a few hours and wandered on the tree trunk; others abandoned the nest site after a take-over by another male, or after the nest was partially destroyed by rain or the attack of ants.

PAGE 104

96 Guarding females were not common. I observed this behavior only three times during the 1987 season and seven times during the 1988 season. If there were several females on the tree, the one who guarded was the one of longest residency (Table 4-2). All females observed guarding had previously copulated and laid eggs in those nests and had at least one month of prior residency (x=58.7 d, range=30-102 ) . Nests guarded by females did not accumulate more eggs than the population mean (population x=21.44 eggs, sample x ))=24.7±4.19 eggs, t=0.78, p=0.45, two-tailed t-Test) , but had higher hatching success than the population mean (population x=18.47 juveniles, sample x=43.1-11.76 juveniles, t=2.09, p=0.06, two-tailed t-Test). I did not observe females cleaning or repairing nests; rather, they sat inside the nests and prevented the nests from being occupied by others. None of these nests were taken-over nor occupied during the time a female was sitting inside. The females continued guarding even after all eggs had disappeared. No juveniles hatched from those nests while the female was guarding and I do not have evidence for whether or not females actually cared for the eggs. None of the original nest owners returned to the nests. Neither could I determine if females were eating eggs, hence, I do not know if females were at nests securing a food resource. Males can recognize their own eggs and feed only on foreign eggs (Mora 1987). There is no reason to believe that

PAGE 105

97 this ability is exclusive to males. If females of longer residency monopolize copulations, most eggs in a nest would have been mothered by those females. Hence, nest guarding would bring greater benefits to the female of longest residency than to any other female. Female Associations and Nest /Male Measurements Female Associations and Nest Longevity Old nests attracted more females than new nests. In 1988, nests at Sites A and B from the 1987 nesting season {n=18. Chapter 3) had a significantly higher number of females associated with the nest site than new nests (sample x=10.13, t=12.7, p<0.001, two-tailed t-Test) . There were also more resident females at old nests than at new nests (sample x=5.95 females, t=16.8, p< 0.001, two-tailed t-Test). Female Associations and Nest Take-overs For all sites except the Log and the A Sites (A Site removed from analysis for missing data), there was a positive correlation between the number of owners a nest had and the number of females arriving at the nest site during the season (r=0.31, n=101, p< 0.05). A positive correlation was also found between the number of owners and the number of females

PAGE 106

98 associated with the nest for more than 3 d (r=0.22, n=101, p< 0.05). Nests for which males compete more (the ones with higher rates of take-overs) are the ones that females visit more often and to which they remain associated for a longer time (R^=0.08, F=5.37, p<0.05; Multiple regression). Females associated preferentially with nests that had multiple owners. A comparison of the patterns of female associations between the 14 nests of this study that had three or more male residents (Chapter 3) and a random sample of 14 nests with one or two male residents revealed that nests with more owners had significantly higher total numbers of females associated with the nest (U=68, U crit=64, p<0.05; Mann-Whitney U Test) , and higher numbers of resident females (U=70.5, U crit=64, p<0.05; Figure 4-7). The patterns described above could be produced in different ways. It may be that males compete for certain nesting sites to which females are attracted; or that females are attracted to good, safe nesting sites for which males compete. In either case females may be potentially selecting vigorous males. That is, we cannot know if it was the male or the female preference for certain nest sites what produced this pattern. Correlational analyses do not allow us to distinguish between those two explanations but the data suggest that both males and females are attracted to the same nesting sites.

PAGE 107

99 Female Associations and Male Residency Females visited more nests with long male tenures. There was a positive correlation between the length of male (nest owner) residency and the total number of females that arrived at the nest site (r=0.34, n=101, p<0.001). Also, the length of male tenure was associated with the number of resident females (r=0.37, n=101, p=0.001) through the nesting season. This suggests that the same nests consistentlyexperienced long associations of both males and females. Compone nts of Female Success Female Success and Nest Success The fitness of females of ^. albomarainis can be estimated by the number of juveniles hatched at the nest where they oviposited. Estimation of female fitness from the nest hatching success will be necessarily inaccurate because not all eggs that hatched in a nest were laid by a single female. This means that I will overestimate female fitness for nests hatching many eggs that were oviposited by several females. Despite that, we can gain insight to the fitness outcomes of female decisions of staying and ovipositing at a nest site by investigating the correlation between the extent of female residency at nest sites with the success of the nest .

PAGE 108

100 Po pulation patterns . Higher numbers of females were associated with nests that produced the highest number of juveniles in the season (Spearman Rank Correlation, r=0.3, n=101, t=3.5, p<0.05). Also there was a significant positive correlation between the number of resident females (females at sites for more than three days) and the number of juveniles produced at the nest (Spearman Rank Correlation, r=0.28, n=101, t=2.88, p< 0.05). No correlation was found between the number of females overlapping at sites and the number of eggs hatched (Spearman Rank Correlation, r=-0.02, n=101, t=0.48, p>0.05) . Successfu l versus Unsuccessful Males Methods . I compared patterns of female residency status at successful and unsuccessful nests from the pool of nest and male biographies constructed for the 1988 breeding season. I categorized successful males as those with hatching success of more than 20% (x+ SD for the population) and two or more resident females (Chapter 3). There were 13 males clearly successful according to those criteria and they were compared to a random sample of 13 males that had hatching success of less than 20% and did not have more than one resident female. I compared the maximum number of females that ever overlapped at a site, the total number of females remaining at the site for more than three days, and the total number of females that associated with those nest sites between the two groups.

PAGE 109

101 Results . Successful nests had more females overlapping (U=169, U crit=124, p<0.001, one-tailed Mann-Whitney U Test) and more total number of females ever associated with the site (U=169, p<0.001) than unsuccessful nests (Figure 4-6) . Also, females stayed longer (U=167, p<0.001) at successful than at unsuccessful nests. Discussion Not only were more females associated but also they spent more time with successful than with unsuccessful nests. Females arrived at the nest sites at random and they may have chosen to stay at a nest site whenever they found a good nest. I cannot know at this point if the success of a nest/male is the result of female decisions to stay or not. The female decisions to stay and oviposite in a given nest site may influence their reproductive success and consequently, the reproductive success of the resident male. Ovioosi tion Rates The females' decision to stay or leave a nest site will be influenced by the difference between the gain from staying and the benefits a female may obtain from leaving. What are the differences in oviposition success among females that stay at nests as compared with those of females that do not? There are several limitations to consider when addressing the above question. Firstly, it is easier to monitor resident than wandering females; secondly, several

PAGE 110

102 females lay eggs throughout the same period in the same nest, making it almost impossible to attribute eggs to specific females; thirdly, the number of eggs that females lay is not fixed; and lastly, males and females constantly eat eggs from the nests. Methods . One way in which one can attempt to quantify the benefits of residency at a nest site for a female, is to compare the rates of oviposition attained by resident and non-resident females. A reasonable estimate of these can be obtained by removing the resident female and comparing the rates of egg accumulation in a resident's nest with and without the resident present. This will compare the rates at which she oviposited to the rate at which her nest accumulated eggs. I assumed that the presence of a resident female did not discourage other ovipositions . During mid-breeding season in 1988, I selected 16 isolated nests (one nest per tree; sample size limited by availability of nests) that had only one resident female. I removed the female for 10 d and compared the rate of egg accumulation before and after her removal. Results. The nests did not accumulate more eggs before the removal of the resident female than after (Wilcoxon Matched-pairs Signed-ranks Test, Z=1.15, p=0.25; Figure 4-8). Discyssipn. Removing the resident female made no significant difference in the rate of egg accumulation at these nests. These results showed that the rate of oviposition in a nest may be independent of the presence of a

PAGE 111

103 resident female. I have shown earlier that nests with resident females were more successful. Since eggs kept accumulating after removal of the resident female, females may have been staying with good nests rather than the success of a nest being the result of her presence. From a male's perspective, there is very little added value, in terms of egg accumulation, from association of a female. From the female perspective, however, one could argue that nests were accumulating eggs at a slightly higher rate {q^. 2 more eggs. Figure 4-8) before the removal of the resident female and this may give resident females some advantage on a long-term basis if we consider that the breeding season lasts six months and females may increase the difference observed at least 20 times by the end of the breeding season. Discussion: Females Ass ociation with Nests Variation in reproductive success of females is generated by the differences in her survival to reproductive age, in fecundity and mating success, and the survival of the eggs she oviposited at nests. I have related the success of females to the patterns of their associations with males/nests and the hatching success of the nests in which they oviposited. Although such a measure of success only considers the post-mating success and does not take into account the proportion of females that do not reach

PAGE 112

104 reproductive age and do not associate with nests for extended periods (thus underestimating female reproductive success) , it gives information about the adaptiveness of female behaviors and the selective pressures acting on both males and females. The general patterns of female behavior and female success suggest that good nests/nesting sites are an important resource for females. Females depend on nests for safe oviposition sites (they will not oviposit outside a nest), and secondarily as a food resource. In addition, the egg hatching success was found to be more influenced by the nest/nesting site than by the identity of the resident male (Chapter 3). Therefore, the discussion of female behavior and reproductive success is, in this system, a discussion about females association with nests/males and the patterns of success derived from these associations. Not all females in the population were associated with nests at all times during the breeding season. I do not know the proportion of wandering females in the population. These "wandering" females had fewer mature eggs than the average female that visited nests (Figure 4-1) . They could have been younger or older females, or mature females that were ripening eggs away from the nest sites. If the latter is true, why females choose to do so is a question that deserves consideration in future studies. There is also the possibility that those wandering females were excluded from nest sites by means of male rejection or by aggressive

PAGE 113

105 femalefemale interactions. To evaluate these explanations, future studies should measure the rates of survivorship and egg maturation at, and away from, nest sites, as well as the intra-and inter-sexual interactions at the nest sites. The number of females arriving at nest sites varied greatly. The distribution of the number of females visiting sites was found not to be different from that expected by chance. Hence, female arrival at nest sites was due to random encounters with trees and not due to female preferences. Female mate choice, then, probably occurs after the female arrived at a tree. Evidence for this comes from the evaluation of oviposition rates before and after the removal of the resident female. This experiment showed that a good nest will continue to be a good nest despite the removal of its resident female. Also, the total number of females that arrive at the nest site, the number of females that stay for more than 3 d, the total residency time at the nest sites and the total number of females overlapping at sites are all correlated with the success of the nest. Females also had stronger associations with nests that have had more owners, that is, nests for which males competed. Since female visits to trees are random, the pattern that emerges is that they seem to choose to stay at good nest sites . Not surprisingly, the total number of females at sites (an estimation of number of females that visited the nest site) was correlated with the number of resident females (an

PAGE 114

106 estimation of the number of females that stayed and oviposited at the nest at least once; Figure 4-5) . Consequently the likelihood of a male acquiring eggs is influenced by the number of females arriving randomly at the nest site. This may explain in part, why males remained at a nest even when they were receiving none or very few eggs for an extended period of time. Given the mean residency time of females at trees, an average female that survived for the 6 -month nesting season could have potentially associated with five different nests. The difficulty of following females after they left the nesting sites at the forest study sites does not allow me to evaluate in how many different nests an average female oviposited. The long-term association of females with nests produced the overlapping of females at a nest site. This occasionally generated aggressive female-female interactions. These interactions of females were not more common at more successful nests. However, I suggest that these interactions may serve to establish dominance hierarchies that allow females of longer residency to monopolize most of the copulations at that nest. Females may be selected to compete if males and/or nests are limited. Three reasons why males and nests may be limited resources could be that (1) space in nests is limited, (2) sperm is limited or (3) there is a limit to the number of eggs a male can guard. I have no evidence supporting any of the above hypotheses. Testing for the validity of these ideas could be an area of interest for

PAGE 115

107 future studies. Females establish dominance relationships with other females in many other species. The usual outcome of female hierarchies is high ranking females monopolizing most of the matings or achieving higher offspring survival. This has been thoroughly documented in many species of primates (Altmann ai, 1988; Cheney , 1988) . Competition among females for access to mates and/or resources associated with males is not an uncommon outcome for role-reversed systems (Petrie 1983). In red-winged blackbirds, consensus among females in their territory choice is thought to result in the overlapping of females within high quality territories. It has been recognized that within territories of a given quality, a female's reproductive success is reduced in relation to the number of females present, causing competition among females (Petrie 1983). It is possible that this same level of competition among females is occurring in Z. albomarainis . This idea will be explored further in Chapter 7 . This study reports for the first time females "guarding" nests after males had left the site. Females observed guarding were the ones of longest residency and they guarded the nests with higher hatching success. One reason that it may be advantageous for a female to remain at a nest site, is that may help to prevent the loss of her eggs to predation. By staying at a nest site for an extended period, she may establish dominance over the other females and she may

PAGE 116

108 increase the hatching success of her eggs by sitting inside the male-less nests. A surprising result was that egg accumulation in a nest did not differ after a resident female was removed. This suggests that a male gains very little from the addition of a resident female. For females however, staying at a nest site may be adaptive as they may decrease the predation rate on the eggs and the risks associated with searching for another mate, and they may be able to monopolize matings by excluding other females. To analyze the trade-offs of female behavior is a challenging task. I cannot entirely separate the male and the nest effect influencing female decisions using correlational analyses. Another problem is the question whether nests were more successful because they got and retained more females, or did they in fact attracted more females because they have something (a male or a nest) that females were looking for? The patterns of female residency, the oviposition rates at nests with and without female residents and the patterns of nest success all suggest that the underlying cause of the observed patterns was that females were associating with successful nests. In summary, I have shown the patterns of female associations to nests and the reproductive outcomes of those associations. I argue that females arrived at the nest sites randomly and stayed at good nesting sites. I hypothesize that there was competition among females for access to good

PAGE 117

109 nesting sites as reflected in aggressive femalefemale interactions. These interactions were not common, but would prevent the access of some females to matings with a resident male. Females of longest residency at the nests may have an advantage over females of shorter residency. However, the question of what females are actually looking for when making mating decisions is unanswered. Do females associate, and stay at nest sites because of the male holding the nest or because of the nest itself? How important is the male for female decisions of staying at a nest site? These questions will be evaluated in Chapter 5.

PAGE 118

110 Table 4-1. Sex ratios (males : females) for the two breeding seasons studied. Year Site Sex ratio Total sample size 1987 Log 1:3 165 A 1:6 418 B 1:2 65 Total 1:4 648 1988 A 1:3 158 B 1:4 82 C 1:2.5 653 Total 1:3 893

PAGE 119

Ill (U 4.) (0 03 CO CO CO 4J (U CO d o ro in cn in CN ro rH m CN IT) in rO rH c^ ro iH » « CN O o ro ro VD iH n ix> O rO (N tH CN tH o 00 in 00 VD CN O 1> VD CN O in CN VD in VD iH 00 00 u CO d 0 0 d as 0) 0) CO 4-) 4-) d to -rH i-> 4J CO d Q) cu d CO CD u u a -H 4J CO d (U .—1 fO 0) g OJ 4-) 4.) 4H CO 0) u d 0 0 * 4J *

PAGE 120

112 In nest >1 h (n=13) Q Finished laying (n=8) m Bitten-chased (n=6) M In-out (n=7) Wandering {n=12) Type of Nest Visit CQ 01 N Q> 8 1 6 Figure 4-1. Mean number of mature eggs (SE) found in females performing different kinds of nest visits.

PAGE 121

113 a V u PL, 30 (0 <0 20 10 o "* 00 o CN VO CO tH ro 1 1 1 1 H H iH CN in 1 1 1 1 1 9\ H H m m i-t H O c« (N Total No. Females Visiting Nest Sites Figure 4-2: Total number of females that visited nests (n=145) at the forest sites. The average nest received visits from five different females during the complete breeding season.

PAGE 122

114 15 n 10 5 I i I 0 Observed Expected r 0123456789 No. Female visits/30 d Figure 4-3. Frequency of the number of different females visiting 50 nests over a 30 d period and the expected Poisson frequency.

PAGE 123

115 25 1 20 15 10 0123456789 10 11 NO. of Females at Nest Site for more than 3 d Figure 4-4. Number of females remaining at the nest sites (n=145) for more than 3 d for all forest sites and both breeding seasons.

PAGE 124

116 OS « 15 s O a o •H CO 0) Oi 10 0 S5 T T" T 5 10 15 20 25 Total No. Females That Visited Site — I 3 0 Figure 4-5. Number of females remaining at the nest site as a function of the total number of females visited the nest sites (r=0.82, n=145, p<0.001)

PAGE 125

117 35co 30 00 25 in o in o to o in o in o in o • CO 1 10 o CM n in (D 1 1 to 1 1 1 f 1 CO o CM m in lO Maximum Length (d) of Female Residency Figure 4-6. Longest period (d) spent at each different nest site (n=143) by resident females, for all forest sites and both breeding seasons .

PAGE 126

118 Total Total at Site Residents Figure 4-7. Female associations with nests (mean +SE) that experienced three or more resident males through the season as compared to nests that had one or two owners. Sample nests were derived from the C Site. Differences were significant at alpha level of 0.05, Mann-Whitney U-Test.

PAGE 127

119 Figure 4-8. Egg accumulation (mean+SE) at 16 nests 10 d before and 10 d after the removal of their resident females. Nests selected occurred one per tree and had only one resident female. Differences were not significant (Wilcoxon Matched-pairs Signed-ranks Test) .

PAGE 128

CHAPTER 5 DO FEMALES ASSOCIATE WITH MALES OR NESTS? Females arrived at nest sites and remained on trees for variable periods of time. Association of females with the male and his nest has consequences for both female and male reproductive success since longer associations resulted in higher hatching success. I have addressed the problem of distinguishing the male from the nest effect for explaining male success and female behavior. An experiment presented in Chapter 3 showed that the success of a male is stronglyinfluenced by his nest. Here I will evaluate the influence of nests and males on female behavior. Whether females remain at a nest site because of the nest or because of the male is an important question. If female decisions to stay and oviposit at a nest are strongly based on the nests themselves, strong competition among males for "good" nests/nesting sites is expected. Evidence for this competition was given in Chapter 3. This chapter presents an experiment designed to evaluate the question, do females associate with males or nests? One way to answer this question is to determine if females are more likely to leave a nest site when a male was removed. If nest sites were important, females would stay at 120

PAGE 129

121 good sites even after a previously unsuccessful male had been switched to this site The experiment consisted of switching successful and unsuccessful males to good and bad nesting sites. I asked whether resident females remained at the nest sites after the switching or not. It would have been instructive to remove and switch nests. This is practically impossible because nests were built directly on the tree bark and its destruction will follow any attempt at removal. Methods I defined successful males as those with more than one female resident and with more than 20% (above the mean for the population at mid-breeding season. Table 3-12) hatching success. This follows the definition of "good" male employed before. For this experiment however, I chose only males with females associated for more than 9 d (that is, females that potentially oviposited in the same nest at least twice) and had confirmed oviposition associations with the males. Therefore good nest sites were those with successful males. Unsuccessful males were assumed to be at bad sites and were defined as those with no females remaining associated with the nest site for more than nine days and hatching success of less than 20%. This criterion for defining success of nests and males had the advantage that it took into account a male's success at acquiring ovipositions and his parental

PAGE 130

122 abilities but had the disadvantage that it established a dichotomy in nest sites and males that did not take into account the fact that there was a gradation of intermediate success in the population. However, the criterion was useful in separating the very bad nest sites and unsuccessful males from the rest and was the simplest variables I could use to compare female residency behavior when associated with different sites/males. I ran this experiment during the mid-breeding season of 1988. It was the same male switching experiment discussed in Chapter 3 . Data analysis differed in this case in that I was interested in comparing the number of female residents associated with the nest sites rather than with the males. I evaluated the nest censuses for a period of 15 d considering only nest sites with only one nest per tree. I was able to identify a total of 23 nests that clearly met the above criterion: there were 15 successful males at good sites and 18 unsuccessful males at bad sites. A total of seven successful males were switched to good sites and eight were switched to bad sites. All switches were random. Of the 18 unsuccessful sites and males identified, eight males were switched to previously successful sites and 10 were switched to unsuccessful sites. Only one pair of males was exchanged through this random assignment. If by randomization an individual was to remain in its original site (one time in this experiment), I gently picked it up and removed it for four hours (the time it took to complete the switchings)

PAGE 131

123 before returning it to its original nest. Data on the number of females resident on trees before and after the switching for each nest were compared using Wilcoxon Signed-Rank Tests, one-tailed, at an alpha level of 0.05. Regyltg; Female Residency Before and After the Switchings I found no significant differences in the number of females associated with nest sites before and after the switching, regardless of whether the females were at good or bad nest sites. Females associated with good sites did not leave the site after the introduction of a previouslyunsuccessful or a previously successful male and the introduction of a previously successful male did not increase the number of females attracted to bad sites (Table 5-1, Figure 5-1) . DiSgUSSion; Females Associati ons With NpstR I predicted that if the resident male at a nest were not important for a female's decision to remain at a nest site, there should be no changes in female residency after the switching of males. If the success of a nest were due to the identity of the resident male, females should leave good sites when a previously unsuccessful male was introduced, and join bad sites when a previously successful male was switched to the site. The results suggest female associations with nest sites were due to their preference for sites where the

PAGE 132

124 nests were built. This was reflected by the lack of change in female residency after the switching of males (Table 5-1) . Female association with nests increases male success for those holding nests at sites preferred by females (Chapter 3). Since sites influence reproductive success of males and females, intense competition for good nest sites should exist among males. Longer residency periods and higher numbers of resident females should be expected at good sites. Supporting evidence is the positive correlation between number of nest owners and nest success and between number of resident females associated with a nest through the breeding season and nest success (Chapters 3 and 4) . But, what is it in the nest sites that produces this effect? This question will be explored in the following chapter.

PAGE 133

125 Table 5-1. Residency of females at nest sites after the switching of previously successful and unsuccessful males. Results from the Wilcoxon Signed-rank Tests, two tailed. Residency of females at Switched males n (oairs) z P Good Sites Successful 8 -0. 45 >0. 05 Unsuccessful 7 -0. 91 >0. 05 Bad Sites Successful 8 0 >0. 05 Unsuccessful 10 -1 >0. 05

PAGE 134

126 n 00 0) 0) S M 0) a « T3 H CO ai 0) 4J 0) 4J At •H o o oa n < 3 1 2 1 8 I Before Switching 0 After Switching 10 G/G G/B B/G B/B Male and Nest Site Figure 5-1. Changes in the mean number of resident females at good and bad nest sites after the switching of previouslysuccessful and unsuccessful males. Residency of females was followed 15 d before and 15 d after the switching. Sample sizes are indicated above the columns. G/G=males from good sites switched to good sites, G/B=males from good sites switched to bad sites, B/G=males from bad sites switched to good sites, B/B=males from bad sites switched to bad sites.

PAGE 135

CHAPTER 6 THE NEST SITES The reproductive success of males is associated with and is indistinguishable from the success of their nests. I have shown this through the experiment in which I switched males to different nests (Chapter 3). In that experiment previously unsuccessful males increased their success when they were switched to nests of previously successful males. The success of a nest may be related to any of two basic elements: the nest itself or the nesting site. Nests and nesting sites may be the key to understanding some features of the mating system of this species. Also, it may be the key to identifying important variables contributing to the variation in reproductive success of both males and females. Male-male competition and female choice seem to show elements of a resource-based mating system (Chapters 3 and 4). From the patterns of nest use by males, of nest and male success and of female behavior two general hypotheses can be derived. (a) There are good and poor nest sites. This is reflected in the higher number of take-overs, the number of different occupants a successful nest experienced through the season and the building of nests in sites of former nests. 127

PAGE 136

128 My observations and experiments on female behavior also suggest that there are good and poor nesting sites. The male switching experiment showed that there are good nest locations that retained resident females regardless of the identity of the male (Figure 5-1). Although females arrive at nest sites randomly, ovipositions were biased towards a few nests (Figure 3-1) . More females associated with "good"sites , i.^., the ones that experienced more take-overs or were re-utilized more heavily (Figure 4-7). (b) Good nesting sites are limited. There were not only differences in the "quality" of nests/nest sites, but also an indication that some resource, whether the nest or the nest site, is limited. Several observations led to this hypothesis. (1) Not all adult males built nests, instead, they took over nests or reused abandoned nests. (2) Females remained associated with nests even after the nests were abandoned or after the male had been switched. Males compete for nests and one reason may be that nesting material is limited. Mud and bark are plentiful. The nest material that is most likely to be limiting to males is the silky secretions males add to build and repair the nests. These secretions may be costly for males. Not all males built nests and only successful males repaired the nest walls (Table 3-14). I have not found evidence for specialized glands producing the silky material. I assume the material used is salivary secretions produced by all

PAGE 137

129 mature males. Aside from this, nesting sites themselves may be limited. If sites are important, what makes a good site? A nest site is a complex microhabitat characterized by a number of physical and biological variables such as the species and architecture of the tree where the nest is located, the microclimate around the nest, and the array of other species utilizing the tree. My approach to determine which variables are important for the differences in success of nest sites was to look at the sources of nest loss and egg mortality . The major sources of egg mortality and nest loss throughout the two years of this study were rain, fungus, ants, and flatworms (Table 6-1) . I will examine in this chapter the differences among sites for each of the above variables and I will evaluate how they may account for the observed variability in male reproductive success. Finally, I will discuss the importance of nest sites for understanding the mating system of these harvestmen. Rainfall Patterns of Precipitation The most important seasonal change on the Barro Colorado Island is rainfall. The biology of many plant and animal

PAGE 138

130 species is strongly influenced by rainfall (Leigh 1982). Breeding of Z. albomarainis occurs during the rainy season during which they are exposed to heavy and unpredictable rains . The island received an annual average of 2686 mm (SE=220.27, n=:10 years, Figure 6-1) of rain from 1979 to 1988. For the two years of this study, the onset of the rainy season was late April in 1987 and mid May in 1988. On the island, the wet season lasts an average of 34 weeks (Rand and Rand 1982; Windsor 1990). For the 1988 nesting season (May-December), October was the wettest month (Figure 6-2), High winds are often associated with storms (Brokaw 1982) and may cause branches to fall, trees to snap and epiphytes to fall bringing down pieces of bark. Treefalls occur most often in the middle of the rainy season, around August -September (Brokaw 1982) . A heavy storm derived from a hurricane offshore can completely wash a nest off a tree trunk (observed 4 times) . In protected sites the effects will be less than in exposed sites where nests might get damaged during a strong storm. Rain may not only damage the nests physically, but water accumulating in the nests may destroy the eggs or make the nest more susceptible to fungal attack. Harvestmen generally oviposit in humid places (Rambla 1973) and it is expected that in this species eggs also require some humidity to develop successfully. However, too wet a site carries the possibility of flooding or structural

PAGE 139

131 damage. Too dry a site may not be adequate for egg development and nest walls tend to collapse when they get too dry. Nest Damage by Rain Rainfall accounted for most of the nest losses (Table 61) . Water running down the tree trunks inundated and washed off nests; and the gusty winds accompanying storms caused branches to fall, producing physical damage to nests. Nest abandonment usually followed big storms. For the 1988 season, 3 6% of all nest abandonment for which I can attribute a cause (11 out of 31 nests abandoned for known reasons) occurred after a storm. Site Differences in Rain Runnin g down the Trees Were some nesting sites significantly wetter than others? Methods . Because water reaching the nests was primarily coming down the sides of the trees, I have evaluated nesting site differences by measuring the amount of rain coming down the sides of the tree trunks. I sampled water accumulated underneath the nests daily for one week in August and one week in September of 1988 from 32 nests at the C Site, 13 nests from the B and 10 nests from the A Site. I collected water in plastic bags (16.5 cm wide x 18.5 cm long) attached to the bark directly underneath nests

PAGE 140

132 (Figure 6-3) . The bags had a metal semi-circular piece at the opening facing the nest. The metal piece held the bag opened and helped to prevent it from collapsing when water and debris accumulated inside. I removed water from each bag with a 10 ml hypodermic syringe (accuracy of -1 ml) . Bags were emptied and debris was removed after each water collection. Samples were taken between 08.00 and 10.00 h. Since most rain falls in the afternoons (Rand and Rand 1982; personal observation) I attributed the rain collected daily to the amount of rainfall of the day before. During the same period in which I sampled water under the nests, male success was measured as the number of eggs hatched at the nest and the number of resident females associated with the nesting site . Results Comparisons between study sites . There were no significant differences in the total amount of rain received by nest sites between one patch in the forest and another (Kruskal-Wallis one-way Anova, H=4.42, d.f.=2, p>0.05). Comparisons between nesting sites . The amount of water coming down the tree trunks was always much below the total rainfall recorded for that day. The variation in rainfall recorded in the bags was greater under heavy than under light rain. In a heavy storm, the wettest sites received five times more rain than the driest sites (Figure 6-4) . The differences among individual sites were consistent. Individual site contribution to the variance was highly

PAGE 141

133 significant (F ratio=variance between groups divided by variance within groups=21 . 65 , d.f.=10, 238; p<0.001). There were then, drier nest sites that consistently received less rain than the wettest sites. Does Rain Influence Male Reproductive Success? I investigated whether the amount of rainfall running down tree trunks influenced the success of nests in three different ways. The first two correlative analyses established the relationship between nest success and the amount of rain at the nesting sites. Lastly, I present an experiment that investigates whether changing the amount of water reaching the nests influences their success. Methods . I ran correlations between the amount of rain and two measures of success, the number of resident females (Chapter 4) and the number of young reared at nests (Chapter 3) for the 55 nests under which I collected rain. I also ranked the 32 nests monitored at Site C according to the amount of rain they accumulated over a one-week period. The 10 highest ranked nests (those that accumulated the most rain) were categorized as wet sites and the 10 lowest ranked nest sites (those that accumulated the least rain) were considered dry sites. The remaining 12 nests were categorized as intermediate sites. I performed one-wayanalysis-of -variance to evaluate the association between water and nest success.

PAGE 142

134 Roof experiments. In addition, I modified the amount of rain running down the tree by placing roofs over nests. In August of 1988, I randomly selected 30 nests from my three study sites. Half were randomly assigned to an experimental group that had a reduced amount of water coming down the trunks. The other half were assigned to a control group. To decrease the amount of water reaching the nests, I placed artificial roofs above the nests; control nests had only a piece of thin wire above their nests. Roofs were semi-circular shaped pieces of rigid plastic measuring 16.5 cm diameter and 12 cm radius. They were attached to the tree bark 10 cm above the nest (Figure 6-5) . In preliminary trials, I confirmed that rain was completely excluded with this procedure. I monitored the number of juveniles hatched at the nest sites and the number of resident females at the sites over a 20 d period before the placement of roofs and compared them to those over 20 d after the placement of the roofs (twotailed Mann-Whitney U Test) . I predicted that a wet nesting site was not a safe site and consequently, decreasing the amount of rain reaching the nesting sites would increase the hatching success at these nests. Results Rain and patterns of success . Drier nests did not have higher success than wetter ones. The total amount of water accumulated underneath the nest sites was not correlated with the number of eggs hatched at the nests (r=0.2, n=55, p=0.3;

PAGE 143

135 Figure 66a) or with the number of resident females associated with the nests (r=0.2, n=55, p=0.2; Figure 6-6b) . When evaluating the success of nests categorized as dry, intermediate and wet, I found that nests receiving intermediate levels of rain had a slightly higher hatching success (Figure 6-7a) , This difference, however, was not significant (Kruskal-Wallis One-way ANOVA, d.f.=2, H=0.08, p>0.05). The driest nests had a larger number of resident females associated with the sites, but the difference was, again, not significant (Kruskal-Wallis One-Way ANOVA, d.f.=2, H=0.84, p>0.05; Figure 6-7b) . The eff ect of roofs on nest success . Shortly after the initiation of the experiment, two nests were lost from the experimental group (one nest was abandoned and the second was attacked by ants) and one nest was lost from the control group (the male abandoned the nest) . At the beginning of the experiment, nests in the experimental and control groups did not show differences in the number of juveniles hatched over 20 d before the placement of the roofs (t=1.57, p=0.13; Figure 6-8) or in the number of resident females (t=0; Figure 6-9) . Decreasing the amount of rain at nest sites did not change the association of females with nest sites. Two of 13 nests with roofs acquired an additional resident female after the placement of the roof, but the difference was not significant (U=92, p=0.34; Figure 6-9). The placement of roofs above the nests did not change the hatching success of

PAGE 144

136 the experimental nests as compared to the controls (U=99, p=0.2; Figure 6-8). Both groups decreased slightly the number of juveniles hatched over the 20-d period after the placement of the roofs. Discussion I did not find evidence for an association between the amount of rain accumulated beneath the nests and the success of the nests. Roofs made no difference in the success of a nest. I had expected an increased hatching success in the group of nests with reduced rainfall reaching the nests and a female response to the changes in rainfall at the nests. None was observed. I may not have left the roofs long enough to detect the effects of rain exclusion. Twenty days, however, allowed me to follow the success of the eggs oviposited before and on the day that I positioned the roof. The Importance of Rainfall : Discussion I found large differences from nest to nest in the amount of rain running down the trees and reaching the nests (Figure 6-4) . The individual differences are consistent, indicating that there are nests that are consistently wet. Intense observations of nests during three breeding seasons led me to predict that the amount of water that reached the nests would be important in explaining the differences in nest success. However, the amount of rain at the sites did

PAGE 145

137 not explain the differences in nest success, measured as hatching success or female residency at the sites. Despite the fact that a direct relationship was not shown, the importance of rain cannot be ignored. Rain, indeed, caused physical damage to nests and may have influenced the behavior of other organisms including predators, prey and pathogens. Nesting occurs during the rainy season; these organisms are then cueing on precipitation for reproductive decisions. However, the use of rainfall as a cue that influences behavior at one level (for example, to start and stop the nesting season) may not be useful at another level (for example, to stay or leave a nest site) especially because rain does not influence the hatching success of males or the female decisions to stay or leave a nest site. The observed effect of rain is limited to physical damage to nests. Unfortunately, I do not know whether sites without nests had more rain than the active nesting sites. When the full range of possible sites is considered, it may be that males are, in fact, choosing some sites over others according to how wet they are. This needs further study.

PAGE 146

138 Fungus Nest Da mage bv Funaus Fungal infestations of nests were an important cause of nest loss. For the total of 158 nests monitored in two breeding seasons, 22 (13.9%) were lost because of fungal infections (Table 6-1). More nests suffered from fungus infestations in the 1988 season than in the 1987 nesting season (X^-9.14, d.f.=l, p=0.003) despite the fact that the 1988 rainy season was not wetter than that of 1987 (Figure 61) . I have not been able to identify the species of fungus attacking the nests. The infestations may involve both pathogenic and non-pathogenic species (J. Cokendolpher , personal communication) . The most probable route of infection is through the cuticle (Samson ^ al, 1988), making males who have suffered injuries particularly susceptible to infection. Males a nd Fungal Infestation Mglg Ipghgyior One major activity of guarding males was cleaning the nests of fungal mycelia. Males removed fungus infestations by eating the mycelia developing on the nest floor and walls (Mora 1987). They turned over the eggs on the floor of the nest while cleaning them. In this way.

PAGE 147

139 males are able to deter the proliferation of fungus in the nest (Mora 1987 ) . Effect of fungus on males. In my three seasons of intense observation of the population of Z.albomarainis in the Barro Colorado Island, I never observed a single event of predation on adults. The sudden death of guarding males in situ (11%, n=158) remains the only identified cases of adult mortality. Those males that were found dead inside their nests showed proliferation of white fungal mycelia out of their bodies within a few days of their death. I proposed earlier (Mora 1990) that males may have died of fungal infections acquired as a result of nest-cleaning activities. I thought males became infected by eating the spores of the fungus. This is probably incorrect since all known pathogenic fungi that attack insects and arachnids seem to infect their hosts though the cuticle (Samson ^ , 1988). Fungus infestati ons and male survival . For the 22 recorded cases of fungus infestation (Table 6-1) , I categorized the fungal infestation as mild or heavy according to my field note descriptions and compared the incidence of male death after the fungal growth. More males lived after a fungus infestation than died. However, more males died after their nests showed a heavy infestation than after a mild one (Table 6-2) .

PAGE 148

140 Effect of Males on Preventing Fungus Infestations Methods . In November of 1988, I located 40 nests outside my study sites. I recorded the number of eggs hatched from the nests for a 20 d period and then randomly assigned 20 nests to a control group and 20 to an experimental group. I traced the perimeter on acetate sheets and looked for fungal growth in the nests. None of the nests monitored showed visible mycelia at the beginning of the experiment. I then removed males from their nests in the experimental group. I screened all nests to prevent takeovers at nests without males. After 10 d I removed the screens and mapped the areas of fungal growth on nests. I compared the frequency of fungus infestation in the two groups with a two-tailed Chi-squared Test. Total areas and areas of fungal growth were measured by copying the acetate traces onto paper and running the total areas and the areas of fungal growth through an area meter (Li-Cor Model LI-300) . Differences in total areas of nests infected were compared with a Mann-Whitney U Test. fiesults, The presence of a male was effective in preventing fungal growth. No control nests (with males) showed any mycelia after 10 d of observations while six out of 20 experimental nests showed fungal growth within the nests. This difference was highly significant (X^=7.06,

PAGE 149

141 d.f.=l, p< 0.01). The mean area covered by fungus after 10 d was 1.83 cm2 (SE=1.02, range=0 . 2-6 . 85 cm^). Nests with fungi were not bigger than nests that had none. The area of nests that developed fungal growth was not different from the area of nests that did (mean area of nests with fungus=21.19 cm^ , SE=0.6, n=6; mean area of nests without fungus=21.09 cm^, SE=1,3, n=34; U=103.5, p> 0.05, Mann -Whitney U Test) . Fungus Infestation and N est Success The occurrence of a fungus infestation was not correlated with the number of juveniles reared at that nest {r=-0.3, p=0.1, n=158) or with the number of females associated with the nest {r=0.09, p=0.6, n=158) . There was, however, an association between the total rain that reached the nests and the occurrence of fungal infestations (r=0.4, p=0.03, n=55) indicating that the wettest nests had significantly higher occurrences of fungal infestations. The success of the nests that showed infestations was not different from the success of nests that did not over a 20 d period prior to the removal of the male (U=61, ni=6, n2=14, p>0.05, Mann-Whitney U-Test) . Nests that showed mycelia growth were probably not infected prior to the beginning of the experiment and fungi were not decreasing the hatching success of those nests.

PAGE 150

142 The Importance of Fungus: Discussion Parental males that engage in cleaning activities may be at risk of dying from fungus infections. Growth and reproduction of most fungus species is stimulated by high humidity conditions (Campbell 1990) ; therefore wet nest sites would have higher probabilities of fungus attacks than drier ones. Rain may be influencing the success of males and nests through its influence on fungal growth. A male nesting in a wet site may be at a disadvantage compared to males in dry sites because the former may suffer more (or heavier) fungus infestations . The role of pathogens affecting the behavior and reproductive success of organisms is poorly understood. Most biologists assume that pathogens are an important constraint affecting organisms but very few actually measure the impact of pathogens in natural populations. The selective pressure of pathogens has been proposed as a factor that may favor the evolution of parental care (Wilson 197 5) . In earwigs, egg mortality to fungal infections is thought to have selected for brooding behavior (Fulton 1924). The role of fungal infestations in nests of Z. albomarainis is still not completely understood and remains an area that merits future exploration. I do not know what fungal species attack the nests (and the males and eggs?) and if indeed the same fungus that develops in nests is the one that grows on dead males. I have been unable to prove the causal effect of fungus

PAGE 151

143 infection and male death. The isolation of the fungus and experimental inoculation of nests and males are necessary to address these questions. Predators Nest Da mage by Predators Ants and flatworms were the principal predators observed over the two seasons of this study and my previous observations of 1985. Their predation on eggs was confirmed by direct observation. Larvae of Coleoptera (three nests) and small nematodes (one nest) were observed inside nests, but their role as predators was not confirmed and their low incidence in nests suggests that they do not constitute a significant factor in egg loss. Ants attacked more nests than did flatworms (Table 6-1) . There were several species of ants involved, including army ants and ants that forage solitarily. The species of flatworm preying on eggs of Z. albomarainis is yet unidentified; it is a light-brown, yellowish species, about 3 . 5 cm long. Unfortunately, to my knowledge there is no information on the natural history of predaceous flatworms in tropical forests. I do not know the capacity for movement of the species of flatworm attacking the harvestmen or the range of its diet, hence their impact can only be evaluated from my data on the frequency of

PAGE 152

144 predation events observed. These are in general rarely observed. There was also predation on eggs by conspecif ics . Both males and females fed on eggs, but the actual number of eggs lost to predation by conspecif ics was not measured. I have assumed in this study that all nests suffered predation by conspecif ics . Male Behavior and Predators Ants . A male is capable of deterring the attack of a single ant. He is able to pick the ant up, lifting it up in his chelicerae and throwing it out of the nest. However, males abandoned nests when a group of ants entered the nest. The damage caused by ants was easily recognized. They removed all eggs from the nest floor, de-compacting the mud of the floor, thus leaving an empty nest with the mud of the floor pulverized. No ant-damaged nests were re-used and all males abandoned the sites a few days following the ant attack. Flatworms . Males were not able to prevent the entrance of flatworms into their nests. Flatworms were quite large compared to nests. They covered about one-fourth of the nest area. Flatworms entered the nests by climbing the walls and remained inside the nests for 4-9 d consuming eggs, moving very little. Males did not leave the nests while the flatworms were inside; instead, they stayed against the wall

PAGE 153

145 opposite to the flatworm. I did not observe any other interactions between flatworms and males. Only two of the flatworms observed eating eggs (n=8) ate all the eggs present . Conspecif ics . There were aggressive interactions between guarding males and other males and females that visit the nests. Guarding males were effective at preventing the loss of eggs to conspecif ics but not completely (Mora 1987). The Importance of Predators: Discussion Predators are an important ecological variable accounting for the loss of eggs and nests in this population. There is no direct association between the occurrence of predatory attacks on nests and the success history of the nests as measured by the number of juveniles hatched at the nests and the number of females associated with the nests. The patterns of ant foraging are very complex and unpredictable; there is basically no area in a tropical rain forest that will not be visited by ants (E. Adams, personal communication) . The probability of any tree being subjected to ant visits is theoretically the same for all This conclusion, however, seems naive since I have not considered other variables related to the nest sites that may increase their attractiveness to predators. Differences in the complexity of the microhabitat (presence of crevices, epiphytes, tree species where the nest is located) and the

PAGE 154

146 behavior of the predator species (range of their diets, range of movements, seasonal behavioral differences, taxisms) may reduce the likelihood of particular nests losing eggs to predators. There may also be ecological conditions that increase the probability that some trees will receive more ant and flatworms visits, such as the presence of other food resources or nesting sites for the predators. Exploring the distribution of ant food resources and how it relates to the distribution and success of nests will be an important next step for understanding the relationship between predators and nesting success in this species. What Acco unts for the Differences in Nest Success? I have explored the contribution of rainfall running down the tree trunks, fungus infestations and the occurrence of ant and flatworm predation on the observed variability in nest success measured as number of juveniles hatching and number of females resident at the nests. In an analysis-ofvariance (SAS, GLM model), rain, fungus, ants and flatworms explained little of the variance in hatching success (F=0.47, d.f.=54, p=0.9) or in number of resident females at nests (F=0.71, d.f .=54, p=0.9) . The associations between the variables I have discussed in this chapter are weak and are presented as partial correlations in Table 6-3. Those correlations represent the correlations between two variables while holding all others

PAGE 155

147 constant, that is, they are based on the variation common to the two variables. The diagonal of the table shows the squared multiple partial correlation values. These values represent the proportion of the variance for that variable that is common with all other variables. It reads, for example, that 31% of the variation in the number of juveniles may be predicted (in a linear regression sense) by the five other variables. The variables considered explain little of the observed variation in nest success. These variables, however, constitute a homogeneous collection of variables suitable for factor analysis (total matrix sampling adequacy=0 . 54 ) . An analysis of principal components by factors (Orthothran/Varimax transformation, Statview™) solves the correlation matrix by the identification of three uncorrelated factors. The first factor relates to the nest success and groups two variables measuring it; 72% of the variance in the number of juveniles and 52% of the variance in the number of resident females are predicted by this factor. The second factor groups the total amount of rain at nests and the occurrence of fungal growth. It explains 73% and 74% of their respective variances. A third factor explains 91% of the variance in the occurrence of predation by flatworms at nests. No factor satisfactorily explains a significant amount of the variance in the occurrence of ant attacks. This analysis reveals that the success of a nest is not explainable by any or a combination

PAGE 156

148 of the variables considered. It confirms the association between the amount of rain a nest receives and fungal growth. These findings do not exclude the biological importance of the variables studied since they are the sources of nest and egg losses observed in three breeding seasons . However they suggest that a different approach to understanding what contributes to the reproductive success of the individuals of this population is needed. My working hypothesis was that there are good and poor nesting sites. I then asked what makes a good nesting site. Neither a single nor a combination of the factors known to contribute to nest and egg losses predicted the success of the nest. The answer is clearly not a simple one: a site too wet, too dry, prone to fungus attack, and so on, does not define what a good site is in terms of number of young reared. One important gap that needs to be filled is how does a tree that has a nest differ from an adjacent one that is not occupied by nesting organisms. Perhaps the key to understanding what is a good site is to understand what are non-nesting sites. The evidence that points to the importance of nesting sites is still strong and many interesting questions keep emerging. Why are some trees utilized consistently year after year? Why are nests concentrated only in certain patches on the island (Figure 2-1)? My results to this point suggest that. perhaps the patterns are derived from ecological pressures of which we know little. It could be that the

PAGE 157

149 observed patterns of nest and male success, as well as of male and female behavior may have been selected for pressures no longer present in the population or that are manisfested sporadically. What I am suggesting is that certain trees may have been better nesting sites in previous seasons, or are consistently better throughout a long period of time and become "historical hot spots". The reutilization of nesting sites is possible because opilionids do not disperse very far and individuals of this population are found in the same general area generation after generation. Such behaviors cannot be explained with information collected in the present because the selective pressures may no longer be present or as intense. Or it could be that ecological and social pressures that selected for this mating system may become important only under conditions not present every year. For example, the important ecological pressures accounting for the evolution of various morphological and behavioral traits in Galapagos finches were not identified until a severe drought occurred (Boag and Grant 1981) . The need for a longterm study becomes apparent if we are to understand what are good sites in this population and how the quality of the sites contribute to the variation in reproductive success.

PAGE 158

150 Table 6-1. Factors accounting for nest failure. Frequencies of fatalities are reported; numbers in parentheses are percentages of the total number of nests monitored. Year Site(n ) ^ Rain Fungus Ants Flatworms Unknown 1987 Log 36 4 (11.1) 5 (13.9) 3 (8.3) 0 2(5.5) A 22 6 (27.3) 1 (4.5) 5 (22.7) 1 ( 4.5) 1(4.5) B 16 5 (31.3) 0 2 (12.5) 0 0 Total 74 15 (20.3) 6 (8.1) 10 (13.5) 1 ( 1.3) 3(4.1) 1988 A 11 0 2 (18.2) 0 1 ( 9.1) 0 B 15 1 (6.7) 3 (20.0) 1 (6.7) 3 (20.0) 1(6.6) C 58 12 (20.6) 11 (19.0) 16 (27.6) 4 ( 6.7) 0 TQt^ l M 13 (15.5) 16 (19.0) 17 (20.2) 8 ( 9.5) 1(1.2) Overall 1^^ 28 (17.7) 22 (13.9) 2 7 (17.1) 9 ( 5.7) 4(2.5) ^(n)= total nests monitored

PAGE 159

151 Table 6-2. Male mortality after fungus infestation. Level of Fungal Infestation Mild Heavy Male Died 2 5 Male Lived 13 2 y?= 7.43, df=l, p<0.01

PAGE 160

^ 152 If. Table 6-3. Partial correlations for four ecological variables related to nesting sites: total rain collected at nests over 15 d (Rain) , the occurrence of fungal growth at the nest (Fung) , the occurrence of ant predation (ants) and of f latworm predation (F-worm) ; and two variables related to nest success: the number of juveniles reared at the nest (Juv) and of females resident at the nest (Fem) . The diagonal represents the squared multiple partial correlation. Total • nests analysed=55. Rain Fung Ants F-worm Juv Rain Fung Ants 0.19 0.27 -0.21 -0.07 -0.06 -0.12 0.19 0.16 0.08 0.18 0.0 2 0.14 F-worm -0.23 0.17 0.03 0.33 -0.12 -0.20 Juv Fem 0.31 0.42 0.21

PAGE 161

153 5000, 78 79 80 81 82 83 84 85 86 87 88 89 Year Figure 6-1. Annual rainfall at the Barro Colorado island for the period 1979-1988. Source: Windsor 1990

PAGE 162

154 500, Month Figure 6-2. Monthly rainfall for 1988. Source: Windsor 1990

PAGE 163

155 18.5 cm o ' ^ metal ring 16.5 cm Figure 6-3. Bags used to collect the rain running down the sides of the trees. Rain was collected daily from 55 nesting sites during one week in August and one week of September, 1988.

PAGE 164

156 o 0) o § n u •H "0 (d a 8 1 6 4 2 — I ' I ' 1 ^ 10 15 20 Total Raln£all (mm) 2 5 3 0 Figure 6-4. Rain (mean,SD) running down the tree trunks collected in bags beneath the nests daily (n=55) as a function of the total rainfall recorded for that day.

PAGE 165

157 Figure 6-5. Roofs utilized to decrease the amount of water reaching the nests, (a) Roofs were made of rigid plastic and placed 10 cm above 15 experimental nests, (b) Detail of the shape and dimensions of the roofs

PAGE 166

158 a 9 > o as 180 160 140 120 100 80 60 40 20 0 T (a) T -r^ — I — I — I — »-*T — i-^T — 0 20 40 60 80 100 120 140 Total Rain at Nast Slta (mm) a a 10 n 8(b) 4•• • O T >T 0 2 0 4 0 6 0 8 0 100 12 0 140 Total Rain at Nest Sites (mm) Figure 6-6. Total rain coming down the tree trunks collected at the nest sites over two weeks and the success of the nests (n=55) (a) measured as the number of juveniles hatched at the nests, (b) measured as the number of resident females associated with the nests.

PAGE 167

159 a IH O 0 as (a) (n=12) (n=10) (n=10) 60 40 20 Wet iBt Dry Nest Site « S o h a o H a o t6 o (b) 5n 4311 I, Wet lilt Dry Nest Site Figure 6-7. The success of nest sites for nests categorized according to the total rain collected beneath the nests over a week (a) measured as the number of juveniles hatched at the nest, (b) measured as the number of resident females associated with the nest.

PAGE 168

160 o « A O *i m H a o > b EXPERIMENTAL (WITH ROOF) N=13 CONTROL (OPEN) N=14 64 2 Before After Figure 6-8. Number of juveniles hatched at nest sites (SE) over a 20 d period before and after the placement of roofs above the experimental nests.

PAGE 169

161 Figure 6-9. Number of resident females (SE) associated with nests over a 20 d period before and after the placement of roofs above the experimental nests.

PAGE 170

CHAPTER 7 SITE-BASED MATING SYSTEM OF ZYaopachylus albomarainis The mating system of a population refers to the general behavioral strategy employed in obtaining mates. Its characterization traditionally includes determining (a) the number of mates acquired, (b) the manner of mate acquisition, (c) the presence and characteristics of pair bonds and (d) the patterns of parental care provided by each sex (Emlen and Oring 1977; Vehrencamp and Bradbury 1984) . Thus, the mating system of a species relates to a large array of behaviors and physical adaptations specific to mating and will include social consequences of these behaviors (Vehrencamp and Bradbury 1984) . Since a mating system includes many different behaviors, it is not enough to account for the presence of a particular behavior, but also for its contribution to an individual's breeding success. Mating systems are very diverse in animals. A single model of sexual selection will not explain all mating systems (Borgia 1979) . Mating systems can be correlated with ecological and with phylogenetic factors (Emlen and Oring 1977) . In particular, the differential ability to control access to mates and resources important for reproduction has resulted in a variety of mating systems that rely on material 162

PAGE 171

163 benefits or certain individual attributes (Maynard Smith 1987) . Most models for the evolution of mating systems suggest that the relatively lower number of gametes produced by females and the generally larger parental investment contributed by them will often result in male-male competition for access to females or resources important to females. These models olso suggest that females should choose males with greater ability to control resources or with higher genetic quality (Orians 19 69; Trivers 1972; Borgia 1979) . A general prediction of most models on the evolution of mating systems is that the extent of parental care by both males and females determines the operation of intraand inter-sexual selection. The consequences of increased male parental investment are varied. When a male gives substantial parental contributions, his parental ability becomes a criterion that females can potentially use to assess male quality (Petrie 1983). Increased male parental investment may free a female to produce more eggs and seek additional mates (Graf en and Sibly 1978) . It may also produce various degrees of sex-role reversal, with an increase in the importance of male mate choice and femalefemale competition for high quality parental males (Trivers 1972) . These predictions have been found to hold true for studies on species with paternal care (Baylis 1981; Petrie 1983). However, there is great variation in the degree of

PAGE 172

164 sex-role reversal, the emphasis of intrasexual competition and the modes of mate choice in mating systems with paternal care. In Z.albomarainis . I have shown that exclusive male egg-guarding behavior has resulted in sex role reversal. There is evidence for the importance of male mate choice. It is reflected by male rejection of courting females. There is also female-female competition for access to mates at nesting sites, which may take the form of aggressive female-female interactions or increased matings by females of longer residency at the nest sites (Chapter 4). The existence of paternal care is an important feature of this mating system and influences the general mating strategy by constraining the course of action that males and females can follow to maximize their mating success. The contributions of ecological, physiological and phylogenetic constraints in shaping the mating systems have been recognized as important, but have not been identified for most species. Remarkable exceptions are the studies on kittiwakes (Thomas and Coulson 1988), dunnocks (Davies 1985), bullfrogs (Howard 1988), tungara frogs (Ryan 1985), odonates (Fincke 1988), and red deer (Clutton-Brock 1982). All were long-term studies that combined extended observations of natural populations with various field or laboratory manipulations. For arachnids, such studies are absent (Thomas and Zeh 1984) , I have characterized the major aspects of the mating system of Z.. albomarainis based on my two-year study of a natural population. In this chapter, I

PAGE 173

165 will summarize the main features of this mating system and will present a hypothetical scenario for its evolution. I argue that both male and female mating strategies are determined by variation in nesting site quality and the opportunity to control this resource, making this a sitebased mating system. The Mating System of Z. albomarainis An adequate analysis of the patterns of mate acquisition should emphasize individuals maximizing their reproductive success (Trivers 1972). As with other mating systems, five factors are necessary to understand this mating system: (1) the number of mates individuals acquired; (2) the opportunities for mating for males and females; (3) the opportunities for males and females to control important resources and mates; (4) the patterns of mate choice; and (5) the components of reproductive success for both males and females . Number of Mates Based on the number of mates acquired, the manner of mate acquisition and the pattern of parental care, the mating system of these harvestmen comes close to that characterized as polygynandry (Vehrencamp and Bradbury 1984). This is a mating system of sequential polyandry, with females giving eggs to a sequence of males, and with males collecting eggs

PAGE 174

166 from several females; males are simultaneously polygynous. In this sense, this system resembles those of many substrate-nesting, egg-guarding fish (Perrone and Zaret 1979) . The accumulation of eggs from different females in male nests occurred through their tenure at one nest or at several nest sites during a breeding season. There were, however, nestless males in this population that achieved zero mates. The number of matings a male secured depended on the quality of his nest-site and the length of his residency at the nest (Chapter 3). Females were polyandrous since they copulated with different males at the same nest site or by visiting other nest sites. For both males and females, the number of mates they could acquire was determined by the quality of the nest with which they associated. Males at good nesting sites received more ovipositions (Figure 3-1) and retained more females (Figures 4-4, 4-5 and 4-6). Since good nest sites are competed for by males and had more male residents through the season (Tables 3-4 and 3-5), females at good sites would mate with more males than females at poor sites. There were more females associated with nests that had multiple occupancy than with nests that had only 1-2 owners (Figure 47) .

PAGE 175

167 O pportunities of Mating Male mating opportunities . Since females rarely guard eggs and are able to produce eggs during the entire breeding season, they are available as mates throughout the season. They are not a limited resource for males and mating opportunities for males depend on their acquisition and tenure at a nest. I have assumed that there are no "sneaking" strategies in males and that the estimated operational sex ratio of this population (3 fem:l male; Table 4-1) is accurate and does not represent a real shortage of males . Males could increase their mating opportunities by associating with good nesting sites. There are three lines of evidence suggesting that the quality of the nest site determined a male's mating opportunities. (a) Higher numbers of females associated with nests that had more than three owners than with nests that had only one or two owners (Figure 4-7). This suggests that the mating opportunities for males will increase at sites preferred by females. (b) Males from poor sites increased their mating opportunities (measured as the number of resident females associated with the nest) when they were switched to good sites (Figure 3-5) . (c) Females associated with nest sites regardless of the male identity (Chapter 5), suggesting again that the mating opportunities for males will increase when at sites preferred

PAGE 176

168 by females. Variation in nest site quality and the competition for good sites has long been recognized for other species (Arak 1983). I have reviewed evidence suggesting that good nesting sites are limited in this population (Chapter 6) . Decisions on whether to build a nest in a new site or take-over a nest, whether to abandon or to continue occupying a nest will influence the male's mating opportunities. The benefit for a male to nest in a good site is an increased number of females associated with the site. The cost of defending a good site may be high for males since competition was increased in those sites, as reflected for the higher turn-over of nest owners. However, the benefit of increased mating opportunities at good nesting sites may have selected for males competing for those sites. Female mating opportunities . Female mating opportunities are also limited to their association with nesting sites. Nests are the only suitable oviposition sites since egg guarding is necessary for protecting the developing eggs (Mora 1987). As a result, female mating opportunities are reduced to mating with nesting males. My observations on three breeding seasons confirm that females do not mate outside of nests. Females court males at the nests, hence their mating success will be influenced by their ability to successfully court the nest owner. Since males compete for good nesting sites and as a consequence those sites had a higher number of resident males, female mating opportunities

PAGE 177

169 could be increased by their residence at good nesting sites. Those sites, however, attracted a higher number of females (Figure 4-6) and their mating opportunities at those sites will probably depend on their ability to exclude other females. However, this reduced opportunity for mating at good nesting sites is compensated by the benefits of associating with a high quality territory. The female mating opportunities could also be increased by visiting many different nests. As with males, an important reproductive decision open to females is whether to remain at a nest site or to leave in search of mates and oviposition sites. Whether a female remains at a good nesting site or she visits several nests, she will still probably mate with several different males throughout the breeding season. The costs of searching for a new nest and the probability of finding a better nest once a female leaves a nest site are not known. This imposes limitations to our understanding of the actual mating opportunities for females. Opportunity for Controlling Resources and Mates Nest sites and nests are the critical resource in this mating system. The distribution of good nesting sites is not known and the ecological conditions that determine the quality of a nesting site are not well understood. The following evidence suggests, however, that there is variation in site quality and that good sites are limited. (a) Nests

PAGE 178

170 were concentrated in the central plateau of the island, despite the fact that areas of similar forest composition were available (Figure 2-1) . (b) Oviposition rates were biased towards a few nests (Figure 3-1) . (c) Males took-over nests and reused abandoned nests (Table 3-4) . (d) Females remained associated with the nests even after the nests were abandoned or after the male had been switched (Figure 5-1) . Since females associated preferentially with good nesting sites, males had access to females in a higher number and during longer periods by holding on to a good nest site (Figures 4-5 and 4-6) . Males were able to control and monopolize females through acquisition and retention of good nests . There is also an opportunity for females to control the access of other females to mating and oviposition sites by excluding them from the nesting sites. I have no direct evidence that exclusion takes place, but females engaged in aggressive female-female interactions that may provide the opportunity and the basis for the control of nests and males. I predicted that the length of female residency at the nest site may be one way female dominance is effected. Manipulations of the number and identity of females at nest sites would demonstrate whether some females can control the access of other females to mates.

PAGE 179

171 Mate Choice Male mate choice . Male mate choice may be important in this system, since males reject some courting females (Chapter 4) . It not clear why males reject females, especially because rejection was observed by males with few or no eggs (Mora 1987 and this study) . I hypothesized that rejections were based on the fecundity of females. Rejected females were expected to have no mature eggs and possibly visited nests to feed on eggs. Dissections of rejected females did not support this idea (Figure 4-1) . Since a male can increase his reproductive success by reducing the risk of egg predation by conspecif ics, I predict that male mate choice is partly based on the female's ability to exclude conspecif ics . Body size, age, and length of association with the nest site may reflect the female's ability to dominate aggressive interactions with conspecifics as well as her physiological condition and genetic attributes and may be cues males use for mate choice. Female mate choice . Female mate choice in these harvestmen was determined by the quality of the nest and nest site (Chapter 5, and reviewed in previous section) . This does not exclude the possibility that male phenotypic and genotypic attributes are chosen by females, but their importance probably follows the decision of remaining or not at a nest site. Experimental switching of the resident males

PAGE 180

172 supports the idea that females associate with the nest/nest site rather than with the males (Chapter 5) . The selection of a nesting site does not completely explain female mating patterns. There are male attributes that explain some of the variation in male and nest success. The same males at different nest sites were consistently successful from one year to another and bigger males were slightly (but not statistically significant) more successful than smaller males (Chapter 4) . Guarding abilities or competitive advantage of males may increase with age. Older males acquire and retain good nests more than younger males. Females would be selected to choose mates based not only on the hatching probability and juvenile survival of her eggs, but also on the genetic quality of her offspring (Maynard Smith 1987) , Older males have demonstrated their survival ability, which may reflect their genetic quality. Components of Re productive Success The selective pressures that affect the reproductive success of both males and females are not necessarily the same for both sexes or constant for an individual throughout its lifespan or during a breeding season (Clutton-Brock 1982) . For example, since both males and females can survive through at least two breeding seasons, it is possible that the mating opportunities and reproductive options pursued may

PAGE 181

173 change with age; but the effect of age on breeding success may be opposite for males and females. In Zlalbomarainis . reproductive success of males and females can be estimated as their ability to accumulate eggs at nests and the hatching success of those eggs. For all organisms, the major components of their reproductive success are the individual survival to breeding age, the reproductive life span, the mating success and the offspring survival (Clutton-Brock 1988) . For these harvestmen, several phenotypical and ecological factors such as age, size, physiological condition, fungus infestations or quality of the nest site are important elements of one or several of the above components of success. Following, I will summarize some of the components of breeding success identified in this study. This discussion is based only on factors related to the breeding season of this organisms. Male Reproductive Success Reproductive life span. Males can survive for the complete breeding season and to at least a second season. As such, the population of reproductive males is composed of individuals of different ages. There is a possibility that, as occurs with other species (Howard 1988; McCleery and Perrins 1988), the reproductive options of males change with age. Age may be correlated with the ability of males to build, take-over or hold nests. My limited records of males that survived two breeding seasons {n=18. Chapter 3) suggest

PAGE 182

174 that the breeding success of males did not change from one season to another. Inexperienced males may be at a disadvantage at building nests, preventing the attack of predators or holding nests, but I have no evidence for this. Experimental manipulations and mate choice experiments involving males of different ages, as well as more extensive population studies are necessary to assess whether and how male success changes throughout their reproductive lifespan. Mating success . Male mating success depends on a male's ability to hold a nest at a good nesting site. Body size may also be an important character that determines mate success; larger males had slightly higher success than smaller ones (Chapter 4) . Even if body size is not directly correlated with mating success, still larger males may have a competitive advantage over smaller ones as has been found for other arthropod species (Elgar and Pierce, 1984) . Offspring survival . The guarding abilities of males and the quality of the nest sites influences the hatching success of the eggs. Juveniles leave the nests 2-3 days after hatching and their subadult lives are a mystery. However, juveniles are quite delicate at the time of leaving the nest and their survival will probably be determined by the food availability and the risk of predation at the tree where they hatched. Some parental behaviors of males may improve the hatching success at his nest. Wall repair was correlated with higher hatching success (Tables 3-9 and 3-14), and the

PAGE 183

175 male's ability to prevent the attack of fungus and predators may also increase the success of his offspring (Chapter 6) . Female R eproductive Success Re productive lifespan. Females also had long reproductive lifespans. As in males, the population of breeding females was composed of females of different ages. Whether fecundity, aggressive behavior, or the females' capacity to control nests changes over their reproductive lifespan is not known. If such changes occur, the reproductive options available to females may change. I have no reason to predict changes one direction or another, but studies in other species have shown that the ability of females to control social interactions may increase with age. Fecundity and mate success . Since females mature a few eggs at a time throughout the season, at any one time, there is variation in the number of eggs that the breeding females had available for fertilization. Females have the possibility of increasing or decreasing the number of eggs they are able to mature and of accelerating the rate at which mature eggs are produced. The ability to increase the number and rate of egg production is probably greater at sites with abundant food resources. However, greater competition for those sites and greater costs of excluding other females are also expected at those high quality sites. The comparisons of oviposition rates at nests with one resident female before and after the removal of the female suggest that the

PAGE 184

176 oviposition rates of resident and wandering females are not significantly different (Figure 4-8) . This means that female residency at a nest site may not affect her capacity to increase her production of eggs. It also suggests that there is competition among females for nests. Offspring survival . Although males maintain nests and guard eggs, females make a direct contribution to increasing the survival of their offspring. Resident females may help repel predatory visits by conspecif ics , and may, although rarely, contribute to guarding (Table 4-2). Females will more directly increase the survival probabilities of their offspring by ovipositing at safe sites and by mating with males of superior parental abilities. My data suggest that females choose in this manner. The number and the residency time of female associations with nests were high at nests of high hatching success (Chapter 4) . The Evolution of t he Site-based Mating System of Z. albomarginis . I have argued that the mating system of these harvestmen is site-based. With this I imply that the variance in the quality of nesting sites is the major element contributing to the variance in the breeding success of males and females and that good nesting sites are a limited resource for which both males and females compete. This is not an unusual mating system. In fact, it resembles that of many nesting fish

PAGE 185

177 (Perrone and Zaret 1979) and ratites (Bruning 1974). However, it is unique among the arachnids. There are in general two major evolutionary factors recognized as important in shaping a mating system. On the one hand, ecological factors determine the degree to which mates can be defended or monopolized. These factors include the spatial distribution of resources, the temporal distribution of mates, and some demographic aspects of the population such as operational sex ratios (Emlen and Oring 1977). On the other hand, phylogenetic factors may constrain the ability of individuals to monopolize resources or mates. Among these phylogenetic factors are developmental patterns of the species that influence the need for extended parental care, physiological traits limiting lifespan and multiple reproduction, and life-history traits limiting the ability of individuals to capitalize on the environmental potential for acquiring and monopolizing mates (Wilson 197 5) . I believe the data presented in this study and previous work (Mora 1987) are consistent with the following hypothetical scenario for the evolution of this mating system. There was variation in the quality of oviposition sites and safe oviposition sites were limited. Females were attracted to these safe oviposition sites, which males began to defend against one another to gain access to mates. Territorial males that remained at the oviposition sites and defended them for prolonged periods increased their reproductive success because they acquired a greater number

PAGE 186

178 of mates than males with random or limited access to oviposition sites. They also hatched a greater number of juveniles by (indirectly) defending eggs against predation. Paternal care, thus, is proposed to have evolved from a male mating strategy and it was favoured because egg guarding increased the hatching success of the eggs without decreasing the male's opportunities for obtaining mates. The same basic criteria that females used for choosing safe oviposition sites were then utilized for mate choice. The utilization of these criteria was adaptive for them since males of higher phenotypic and genetic quality were more successful at competing for and defending nesting sites to which females were attracted. Female parental care is the rule in the arachnids. Female spiders (Fink 1984), scorpions (Cloudsley-Thompson 1968), Solifugae (Muma 1966). pseudoscorpions (CloudsleyThompson 1968), Uropygi (Klingel 1963), Schizomida (Cloudsley-Thompson 1968), and Amblypygi (Cloudsley-Thompson 1968) guard eggs and juveniles. There is no case of biparental care reported in this group. There are few other arthropod species with paternal care (Zeh and Smith 1985) . Males guard eggs in assassin bugs Rhinocoris albopilosus (Odhiambo 1959, 1960), E. albopunctatus and a Zelus species (Ralston 1977) and belostomatid water bugs (Smith 1976 a, b) . Male egg guarding in Z. albomarainis was probably derived from no care. Paternal care has been proposed to have evolved from biparental care in most models (Perrone and

PAGE 187

179 Zaret 1979; Werren ^ sX. , 1980), but it also may have evolved from no care in some species of fish (Baylis 1981) . Similar scenarios to the one I propose have been presented to explain other mating systems. Most of them have been developed to explain evolution of the mating systems of fish with paternal care (Perrone and Zaret 1979; Werren ^ al . . 1980) . Several natural history traits of males and females, common to other opilionids, could be viewed as preadaptations for the evolution of this mating system. The long reproductive lifespan of these organisms makes it possible for males to defend territories for extended periods. This is because females are reproductive throughout the complete breeding season. Since females produce eggs continuously, there are no periods in the breeding season in which the cost for males of defending a nest is greater than the mating benefits derived from it. Also, because eggs are very susceptible to predation and fungus attacks, egg guarding is highly adaptive and derives benefits to both males and females . It is indeed unfortunate that so little is known about the natural history and ecology of natural populations of harvestmen and that the phylogenetic relationships among groups have not been fully explored. I have at this point more questions than answers. Many aspects of this system are very mysterious, such as the evolution of the mud nests, the existence of a population of nestless males, and the reasons

PAGE 188

180 for male mate rejections. Resolving these issues may prove very challenging. Other aspects, such as the clarification of the effect of age on reproductive success, the mechanisms for mate choice, and the identification of what a good (and a bad) nesting site is, represent some of the specific areas that could be explored. An obvious weakness of this study is the lack of data I have accumulated on both nestless individuals, wandering females and habitat heterogeneity. An important goal for future studies should be learning about who are the non-nesting individuals and how do trees with no nests differ from those that have nests.

PAGE 189

LITERATURE CITED Altmann, J, G. Hausfater and S. A. Altmann. 1988. Determinants of reproductive success in savannah baboons, Pa pio cynocephalus . In: Re productive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 403-418. Arak, A. 1983. Male-male competition and mate choice in anuran amphibians. In: Mate Choice . (Ed. by P. Bateson) . Cambridge: Cambridge Univ. Press, pp 181-210. Baylis, J. R. 1981. The evolution of parental care in fishes, with reference to Darwin's rule of male sexual selection. Environ. Biol. Fish. . 6: 223-251. Berland, L. 1949. Ordre des Opiliones. In: Traits de Zooloaie. Anatomie, Syst6matique . Bioloaie . (Ed. by P. Grass^) . Tome VI. Paris: Masson. pp: 761-793. Boag, P. T. and P. R. Grant. 1981. Intense natural selection in a population of Darwin's finches (Geospizinae) in the Galapagos. Science . 214: 82-85. Borgia, G. 1979. Sexual selection and the evolution of mating systems. In: Sexual Selection and Reproductive Competition in Insects . (Ed. by M. S. Blum and N. A. Blum). New York: Academic Press, pp. 19-80. Brokaw, N. V. L. 1982. Treefalls: Frequency, timing, and consequences. In: The Ecology of a Tropical Forest: Seasonal Rhythms and Long-term Changes . (Ed. by E.G. Leigh, A.S Rand and D. Windsor) . Washington, D.C: Smithsonian Institution Press, pp: 101-108. Bruning, D. F. 1974. Social structure and reproductive behaviour in the great rhea. Living Bird . 13: 251-294. 181

PAGE 190

182 Campbell, N. A. 1990. Bioloov . Second Edition. Redwood City, CA: Benjamin/Cummings Publishing Co. Cheney, D. L., R. M. Seyfarth. S. J. Andelman, and P. C. Lee. 1988. Reproductive success in vervet monkeys. In: Reproductive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 384-402. Clousdley-Thompson, J. L. 1968. Spiders. Scorpions. Centipedes and Mites . London: Pergamon Press. Clutton-Brock, T. H. 1988. Introduction. In: Reproductive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 1-6. Clutton-Brock, T. H. , F. E. Guiness, and S. D. Albon. 1982. Red Deer: Behaviour and Ecolocry of Two Sexes . Chicago: Chicago Univ. Press. Coddington, J. A., M. Horner and E. A. Soderstrom. 1990. Mass aggregations in tropical harvestmen (Opiliones, Gagrellidae: Prionostemma sp.). Revue Arachnoloaique , 8 (13) : 213-219. Darwin, C. 1871. The Descent of Man, and Selection in Relation to Sex. London: John Murray. Davies, N. B. 1985. Cooperation and conflict ammong dunnocks. Prunella modularis . in a variable mating system. Anim. Behav . . 33: 628-648. Dietrich, W. E. , D. M. Windsor and T. Dune. 1982. Geology, Climate and Hydrology of BCI. In: The Ecology of a Tropical F orest: Seasonal Rhythms and Long-term Changes . (Ed. by E.G. Leigh, A.S Rand and D. Windsor) . Washington, D.C: Smithsonian Institution Press, pp: 2146. Dunbar, R. I. M. 1983. Life history tactics and alternative strategies of reproduction. In: Mate Choice . (Ed. by P. Bateson) . Cambridge: Cambridge Univ. Press, pp. 423433 .

PAGE 191

183 Edgar, A. L. 1971. Studies on the biology and ecology of Michigan Phalangida (Opiliones) . Misc. Pubis Mus Zool. Univ. Mich . . 144: 1-64. Edgar, A, L.and H. A. Yuam. 1968. Daily locomotory activity in Phalanaium ooilio and seven species of Leiobunum (Arthropoda: Phalangida). Bios . 39(4): 167-176. Elgar, M. A. and N. E. Pierce. 1988. Mating success and fecundity in an ant-tended lycaenid butterfly. In: Reproductive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 59-75. Emlen, S. T. and L. W. Oring. 1977. Ecology, sexual selection and the evolution of mating systems. Science . 197 :215-223 . Fincke, 0. M. 1988. Sources of variation in lifetime reproductive succcess in a nonterritorial damselfly. In: Re productive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 24-43. Fink, L. S. 1984. Maternal guarding behavior in the green lynx spider Peucetia viridans (Hentz) . Master of Science Thesis, University of Florida, 84 pp. Foster, R. and N. V. L. Brokaw. 1982. Structure and history of the vegetation on Barro Colorado Island. In: The Ecology of a Tropical Fore st: Seasonal Rhvthms and Longterm Changes . (Ed. by E.G. Leigh, A.S Rand and D. Windsor) . Washington, D.C: Smithsonian Institution Press, pp: 67-81. Goodnight, C. J. and M. L. Goodnight. 1942. Phalangida from Barro Colorado Island, Canal Zone. Am. Mus. Novit . , 1198: 1-18. Graf en, A. and R. Sibly. 1978. A model of mate desertion. Anim. Behav . . 26: 645-652.

PAGE 192

184 Hillyard, P. D. and J. H. P. Sankey. 1989. Harvestmen . 2nd. Ed. Synopses of the British fauna, new series no. 4. The Linnean Society of London and the Estuarine and Brackish-water Sciences Association. Avon: The Bath Press . Holmberg, R. G. and N. P. D. A. Angerilli. 1984. Overwintering aggregations of Leiobunum paessleri in caves and mines (Arachnida: Opiliones) . J . Arachnol . . 12: 195-204. Howard, R. D. 1988. Reproductive success in two species of anurans. In: Reproductive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 99113 . Juberthie, C. 1965. Donnees sur I'ecologie, le developpement et la reproduction des Opiliones. Rev. Ecol. Biol. Sol .. 2 (3) : 377-396. Juberthie, C. 1967. Siro rubens. Revue. Ecol. Biol. Sol.. 4: 155-171. Juberthie, C. and J-F. Manier. 1976. Les grands traits de la spermiogenese chez les Opilions. C. R. Col. Arachnoloaie Fr . . Les Eyzies . Station Biologique Les Eyzies: Academie de Paris, pp 74-82. Juberthie, C. and A. Munoz-Cuevas . 1971. Sur la ponte de Pachvlus cruinamavidensis (Opilion, Gonyleptidae) . Bull . Soc. h-Hi st. Nat. Toulouse . 107(3-4): 468-474. Klingel, H. 1963. Mating and maternal behavior in Thelvohonus caudatus L. (Pedipalpi, Holopeltidia, Uropygi) . Treubia , 26: 65-70. Leigh, E. G. Jr. 1982. Introduction. In: The Ecology of a Tropical Forest: Seasonal Rhythms and Long-term Changes . (Ed. by E.G. Leigh, A.S Rand and D. Windsor) . Washington, D.C: Smithsonian Institution Press, pp: 1117.

PAGE 193

185 Maynard Smith, J. 1987. Sexual selection: A classification of models. In: Sexual Selection: Testing the Alternatives . (Ed. by J. W. Bradbury and M. B. Andersson. New York: John Wiley, pp. 9-20. McCleery, R. H. and C. M. Perrins. 1988. Lifetime reproductive success of the great tit, Parus manor . In: Reproductive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 136-153. Mora, G. 1987. Male egg-guarding behavior in the neotropical harvestman, Zvaophachvlus albomarainis (Opiliones: Gonyleptidae) . Master of Science Thesis, University of Florida, 92 p. Mora, G. 1990. Paternal care in a neotropical harvestman, Zyaophachylus albomarainis (Arachnida, Opiliones: Gonyleptidae). Anim. Behav . . 39: 582-593. Muma, M. H. 1966. Egg disposition and incubation in Eremobatidae (Arachnida: Solpugida) . Florida Ent . . 49 (1) : 23-31. Odhiambo, T. R. 1959. Pin account of parental care in Rhinocoris albooilosus Signoret (Hemiptera-Heteroptera: Reduviidae) , with notes on its life history. Proc . R. Ent. Soc. Lond. (A)., 34(10-12): 175-185. Odhiambo, T. R. 1960. Parental care in bugs and non-social insects. New Scientist . 8: 449-451. Partridge, L. 1988. Life reproductive success in Drosophila . In: Reproductive S uccess. (Ed. by T. H. Clutton-Brock). Chicago: Chicago Univ. Press, pp. 11-23. Perrone, M. and T. M. Zaret. 1979. Parental care patterns of fishes. Amer . Natur . . 113: 351-361. Petrie, M. 1983. Mate-choice in role-reversed species. In: Mate Choice. (Ed. by P. Bateson) . Cambridge: Cambridge Univ. Press, pp. 176-179.

PAGE 194

186 Phillipson, J. 1959. The seasonal occurrence, life histories and fecundity of harvest-spiders (Phalangida, Arachnida) in the neighbourhood of Durham City. Entomol ogist's mon. Mag . . 95: 134-138. Ralston, J. R. 1977. Egg guarding by male assassin bugs of the genus Zelus (Hemiptera: Reduviidae) . Psvche . 84: 103-107. Rambla, M. 1975. Los Opiliones (Arachnida). I & II Parte. Graelsia . 30: 123-220. Rand, A. S. and W. M. Rand. 1982. Variation in rainfall on Barro Colorado Island. In: The Ecology of a Tropical Forest; Seasonal Rhythms and Long-term Changes . ( Ed . by E.G. Leigh, A.S Rand and D. Windsor) . Washington, D.C: Smithsonian Institution Press, pp: 47-59. Rodriguez, C. A. and S. Guerrero. 1976. La historia natural y el comportamiento de Zygopachylus albomarginis (Chamberlain) (Arachnida, Opiliones: Gonyleptidae) . Biotropica . 8(4): 242-247. Ryan, M. J. 1985. The Tungara Frog . Chicago: Chicago Univ. Press . Samson, R. A., H. C. Evans and J. -P. Latg^. 1988. Atlas of Entomopathogenic Fungi . Berlin: Springer-Verlag . Savory, T. H. 1977. Arachnida . 2nd Ed. London: Academic Press . Smith, R. L. 1976a. Male brooding behavior of the water bug Abedus herberti (Hemiptera: Belostomatidae) . Ann. Entomol. Soc. Amer . . 69: 740-747. Smith, R. L. 1976b. Brooding behavior of a male water bug Belastoma f lumif erum (Hemiptera: Belostomatidae) . J. Kans. Entomol. Soc. Amer.. 49: 333-343.

PAGE 195

187 Stephens, M. 1982. Mate takeover and possible infanticide by a female northern jacana ( Jacana s pinosa ) . Anim. Behav . . 30: 1253-1254. Thomas, C. S. and J. C. Coulson. 1988. Reproductive success of kittiwake gulls, Rissa tridactvla . In: Re productive Success . (Ed. by T. H. Clutton-Brock) . Chicago: Chicago Univ. Press, pp. 251-262 Thomas, R. H. and D. W. Zeh. 1984. Sperm transfer and utilization strategies in Arachnida: ecological and morphological constraints. In: Soerm Competition and the Evolution of Animal Mating Systems (Ed. by R. L. Smith) . Orlando, Florida: Academic Press, pp: 180-221. Thornhill, R. 1979. Male and female sexual selection and the evolution of mating strategies in insects. In: Sexual Selection and Reproductive Competition in Insects. (Ed. by M. S. Blum and N. A. Blum). New York: Academic Press, pp. 81-122. Trivers, R. L. 1972. Parental investment and sexual selection. In: Sexual Selection and the Descent of Man (Ed. by B. Campbell). London: Heineman. pp: 136-179. Werren, J. H., M. R. Gross and R. Shine. 1980. Paternity and the evolution of male parental care. J . theor . Biol . . 82: 619-631. West-Eberhard, M. J. 1979. Sexual selection, social competition and evolution. Proc. Amer. Phil. Soc . 123: 222-234. Williams, G. C. 1975. Sex and Evolution . Princeton, New Jersey: Princeton University Press. Wilson. E. 0. 1975. SociobiolooY Cambridge: Belknap. Windsor, D. M. 1990. Climate and Moisture variability in a Tropical Forest: Lo ng-term Records from Barro Colorado island. Panam^. Smithsonian Contributions to the Earth Sciences, No. 29., Washington, D.C.: Smithsonian Institutiom Press.

PAGE 196

188 Wittenberger, J. F. 1983. Tactics of mate choice. In: M^te Choice . (Ed. by P. Bateson) . Cambridge: Cambridge Univ. Press, pp. 435-447. Zeh, D. W. and R. L. Smith. 1985. Parental investment by terrestrial arthropods. Amer . Zool . . 25: 785-805.

PAGE 197

BIOGRAPHICAL SKETCH Giselle Mora was born in San Jos6, Costa Rica, on 13 September 1958. She developed an interest in biology and arthropods during her early college years. She graduated in biology from the Universidad de Costa Rica in 1982. She worked as Station Manager of La Selva Biological Station, Organization for Tropical Studies (OTS) , as curator at the Museo de Zoologia, Universidad de Costa Rica and as invited faculty in Tropical Biology courses for OTS before entering the Department of Zoology at the University of Florida in 1984. She received her M.S. degree in 1987 with a thesis on egg-guarding behavior of the harvestman Zyaopachvlus albomarainis . She had her first contact with these harvestmen on the Barro Colorado Island in 1983. Her project grew from natural history descriptions, to a study on the patterns and adaptiveness of egg guarding, which in turn led to this study on the mating system and the components of reproductive success of these remarkable harvestmen. She received an Exxon fellowship in 1983, a short-term research fellowship in 1985, and a predoctoral fellowship in 1987 and 1988 from the Smithsonian Tropical Research Institute. She was a teaching and research assistant in the Department of Zoology, University of Florida from 1984-1991. 189

PAGE 198

190 She also received a fellowship award from the Consejo Nacional de Investigaciones Cientificas y Tecnol6gicas (CONICIT) , Costa Rica, to complete her dissertation. She was an honor student at the Universidad de Costa Rica (19781980) , the runner-up for the best student paper award at the American Arachnological Society meetings in 1986, and the chairman of the symposium on breeding systems at the IX International Ethological Conference in Utrecht, The Netherlands in 1989. She is married to Henry M. Bourgeois and upon completion of her Ph.D. will return to Costa Rica to join the faculty of the Escuela de Biologla, Universidad de Costa Rica.

PAGE 199

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. H . ^Jane Brbckmann^^ Chair Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of ('^hyl^sophy . mathan Reiskind, Cochair Associate Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ;rson "Associate Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. J^nh Sivinski Assistant Professor of Entomology and Hematology

PAGE 200

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of->Philosophy . This dissertation was submitted to the Graduate Faculty of the Department of Zoology in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy Associate Professor, Escuela de Biologia Universidad de Costa Rica December, 1991 Dean, Graduate School